Measuring Generativity in Older Adults: The Development of New Scales

Abstract

Erikson already emphasized the importance of remaining generative in old age. Current instruments designed to measure generativity, however, were developed to be used with middle-aged adults. In a sample of older adults (N = 195), three newly developed instruments measuring different facets of generativity, each with several scales, were applied and tested with regard to methodological issues (internal structure, internal consistency of scales), bivariate associations with established instruments of generativity and social desirability, and age and gender differences. The study results confirmed that the theoretically assumed factor structure and characteristics of the scales support the assessment of generativity in a multidimensional way in older adults. Positive associations with the Loyola Generativity Scale (LGS) demonstrate a positive validation result. (PsycINFO Database Record (c) 2012 APA, all rights reserved)

Full text

A.Schoklitsch& U. Baumann: Measuring Generativi ty in Older AdultsGeroPsych24(1) © 2011 Hogrefe Publishing
Full-Length Research Report
Measuring Generativity
in Older Adults
The Development of New Scales
Angela Schoklitsch and Urs Baumann
University of Salzburg, Austria
GeroPsych, 24 (1), 2011, 31–43
DOI 10.1024/1662-9647/a000030
Abstract. Erikson already emphasized the importance of remaining generative in old age. Current instruments designed to measure
generativity, however, were developed to be used with middle-aged adults. In a sample of older adults (N= 195), three newly developed
instruments measuring different facets of generativity, each with several scales, were applied and tested with regard to methodological
issues (internal structure, internal consistency of scales), bivariate associations with established instruments of generativity and social
desirability, and age and gender differences. The study results confirmed that the theoretically assumed factor structure and characteristics
of the scales support the assessment of generativity in a multidimensional way in older adults. Positive associations with the Loyola
Generativity Scale (LGS) demonstrate a positive validation result.
Keywords: generativity, older adult, scale development
Introduction
It was in old age that he saw the fruits of able care, that of
some few things well cared for and of a grand generativity
in which adults show an active if detached concern for life
itself, for grandchildren, and for the wider species (Hoare,
2002; p. 192).
Erikson (1950) introduced the concept of generativity
more than 50 years ago and defined it as “the concern in
establishing and guiding the next generation” (p. 267). He
assumed a developmental model throughout life with
eight stages and defined generativity as the seventh devel-
opmental task in midlife. Although, Erikson initially
thought of generativity as a stage in midlife, he empha-
sized that older people should maintain a dignified gener-
ative function and proposed that grandparenthood offered
individuals a second chance at generativity: “Old people
can and need to maintain a grand-generative function”
(Erikson & Erikson, 1997, p. 63).
The concept of generativity as an important element in
human development, significantly impacting one’s ability
to age successfully, was discussed later by other authors
as well. Fisher (1995) interviewed elderly employees and
found that having a sense of purpose or generativity was
central to their belief that they were aging successfully.
According to Kruse and Wahl (2010), older age offers the
possibility to realize generativity based on acquired ide-
alistic (i.e., experience, knowledge, and time) and materi-
al resources. They describe it as a facet that has yet to be
acknowledged for its significant contributions to aging.
Instruments designed to measure generativity were devel-
oped for use with middle-aged adults. As the majority of
these studies have not examined adults beyond their early
70s, little consideration has been given to whether these as-
sessment methods are suitable or appropriate for older peo-
ple. For example, the Loyola Generativity Scale (LGS; Mc-
Adams & St. Aubin, 1992), the most common instrument
measuring generativity, consists of items that are problematic
or unsuitable for older adults (e.g., regarding current concerns
for having children). Furthermore, past experiences become
more important for older people’s sense of self-concept (e.g.,
Dittmann-Kohli, 1995). Existing scales, however, focus only
on the present, the future, or the recent past, while the newly
developed scales assess generative concerns from both a life
review and a past perspective (parental generative concerns).
The newly developed questionnaires and first empirical re-
sults based on a sample of older adults represent one way of
acknowledging the impact of generativity on the aging pro-
cess and, subsequently, its importance as a topic within the
field of gerontology.
GeroPsych 24 (1) © 2011 Hogrefe Publishing

Erikson’s Concept of Generativity and
Subsequent Developments
Erikson (1950) introduced generativity (vs. stagnation) as
the seventh developmental task in midlife following basic
trust (vs. basic mistrust), autonomy (vs. shame), initiative
(vs. guilt), industry (vs. inferiority), identity (vs. confu-
sion), intimacy (vs. isolation) and preceding integrity (vs.
despair). Accordingly, midlife adults have the responsibil-
ity to bear, nurture, and guide subsequent generations as
well as to develop and maintain societal institutions and
natural resources (Erikson, Erikson, & Kivnick, 1986).
Children are important for generativity but “the mere fact
of having or even wanting children” does not “achieve gen-
erativity” (Erikson, 1950, p. 267). On the other hand, it is
possible to be generative without having children.
The first theorist to expand significantly upon Erikson’s
ideas about generativity was Kotre (1984). He proposed
that four distinct forms of generativity exist: biological, pa-
rental, technical, and cultural. Biological generativity is
about begetting, bearing, and nursing children. Parental
generativity is expressed in feeding, clothing, sheltering,
loving, and disciplining offspring (biological or not) and
initiating them into the family’s traditions. Technical gen-
erativity is accomplished by teachers, who pass on skills to
those less advanced than themselves (e.g., how to read, how
to program a computer, how to perform a healing ritual).
When teachers move from teaching skills to passing on
their meaning, they become culturally generative.
Erikson (1950) always thought of generativity as a mid-
life task. According to Kotre (1996), the schedule for the
appearance is misleading as Erikson failed to differentiate
between the various types of generativity and their rele-
vance. For example, biological generativity – conceiving
and bearing children – has a far earlier onset and conclu-
sion, particularly in the case of women, than cultural gen-
erativity. Empirical support for this suggestion can be
found in Snarey’s (1993) longitudinal study, in which pa-
rental generativity precedes societal (i.e., technical and cul-
tural) generativity. “Societal generativity principally in-
volves caring for other younger adults: Serving as a mentor,
providing leadership, and generally contributing to the
strength and continuity of subsequent generations” (p. 22).
This type of generativity usually begins around midlife and
remains predominant until late adulthood. Similar, Manhei-
mer (1995) proposed that cultural generativity may be
linked to late life “when they have done their begetting,
nurturing and passing on of skills” (p. 17).
Vaillant (2002) proposed that there are six adult life
tasks: identity, intimacy, career consolidation, generativity,
keeper of the meaning, and integrity. Keeper of the mean-
ing implies the role of a wise judge; Erikson assigned parts
of this task to both generativity and integrity. The focus is
on the conservation and preservation of collective prod-
ucts, hence the culture and its institutions rather than the
development of its children. Thus, as suggested by Vaillant,
a 70-year-old is usually better at being a keeper of the
meaning than a 30-year-old.
The specifications of the concept of generativity are espe-
cially important when looking at generativity in older adults as
they suggest that certain aspects become particularly important
in older age (keeper of the meaning, societal generativity).
In the present study Kotre’s (1984) four domains of gen-
erativity were complemented by the addition of social and
ecological generativity. The domain of social generativity
was added as it seemed to be an important aspect of genera-
tivity not yet explicitly discussed in the literature. Some of
the items used were derived and modified from other ques-
tionnaires measuring generativity, e.g., “I enjoy guiding
young people” (Ochse & Plug, 1986). Although, Erikson et
al. (1986) cited the importance of the maintenance of natural
resources to generativity in previous works, this principle has
yet to receive adequate attention. Most of the items for eco-
logical generativity were derived from the principle program
of the Green Party in Austria (e.g., animal protection). This
type of generativity, not included in previous scales, was add-
ed in order to acknowledge that the concept of passing the
environment on to subsequent generations, is of similar im-
portance to the future of humankind, as is, for example, val-
ues and experiences. Given the basic idea of generativity, we
not only need technical and cultural but social skills (key-
word: globalization) and an environment in which to live
(keyword: global warming). In contrast to the other aspects,
ecological generativity requires less contact with other peo-
ple. In sum, from a theoretical perspective, no clear assump-
tions can be made with regards to the strengthof associations
between the different generative aspects.
The Importance of Generativity to
Gerontological Research
According to Erikson and Erikson (1997), the discontinuity
of family life contributes to the lack of that minimum level
of vital involvement that is necessary for staying truly alive
in old age. Erikson and Erikson proposed that “lack of vital
involvement often seems to be the nostalgic theme hidden
in the overt symptoms that bring old people to psychother-
apy. Much of their despair is, in fact, a continuing sense of
stagnation” (p. 63). This suggests that generativity is in-
deed an important factor in maintaining one’s psychologi-
cal health in old age.
P. Baltes and M. Baltes (1990) cite generativity and wis-
dom as examples for a normative definition of an ideal state
in old age. Achieving generativity would then become evi-
dence of a successful aging process. Lang and M. Baltes
(1997) distinguished between three types of generativity in
older age: (1) creating lasting values which includes finaliz-
ing decisions regarding certain life goalsand social contacts,
(2) “keepers of meaning” and thereby improvingthe connec-
tion between change and continuity, (3) self-decentness and
32 A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults
GeroPsych 24 (1) © 2011 Hogrefe Publishing

self-responsibility (e.g., being less demanding and thus re-
lieving subsequent generations). Hagestad and Uhlenberg
(2006) pointed out that age segregation impedes generativity,
particularly the creation and maintenance of a generative so-
ciety. Thus, even if older people want to express their con-
cerns, it is not always possible to achieve generativity.
Today, in Western countries more people are staying ac-
tive and healthy for longer than ever before, and the pro-
portion of older people has increased dramatically over the
last decades. “With more adults living into old age, there
is heightened concern about making the later years of life
quality ones – characterized by good health and positive
well-being” (An & Cooney, 2006, p. 410). A number of
studies investigating the relationship between generativity
and well-being or life satisfaction consistently demonstrat-
ed a positive association (e.g., Ackerman, Zuroff, & Mos-
cowitz, 2000; Keyes, & Ryff, 1998; Ochse & Plug, 1986).
As correlations were especially strong when looking at
generative concerns (e.g., McAdams, St. Aubin, & Logan,
1993; St. Aubin & McAdams, 1995), the newly developed
instruments focus on this facet of generativity. Although
there is no empirical evidence for older age groups, a sim-
ilar association between generativity and positive health
indicators can be assumed.
Objectives
The current study sought to specify and define the generative
concerns of individuals in later life stages and obtain empir-
ical evidence for the three newly developed scales based on
Kotre’s (1984) four age-specific subtypes. Various time per-
spectives (past, life review, present) were introduced, in light
of the increasing levels of importance older adults assign to
the past, in order to see if additional information could be
derived. To validate the new scales, we employed the most
common instrument – the LGS – used to measure generativ-
ity. Former research indicates that it is important to control
for social desirability when investigating generativity (e.g.,
Ochse & Plug, 1986). Therefore,we includeda scale control-
ling for this variable in this study. The influence of gender and
age, variables which were often cited asrelevant when inves-
tigating generativity (e.g., McAdams & St. Aubin, 1992; Mc-
Adams et al., 1993; McKeering & Pakenham, 2000), is also
addressed in the research study.
Method
Participants
Participants were recruited through senior citizen commu-
nity centers. Selection criteria were limited to those aged
60 or older with no cognitive impairment. A total of 195
older adults (110 women and 85 men) ranging in age from
60 to 94 (M= 72.5, SD = 8.8) participated in this study.
Most of them were Austrian citizens (84.6%) living in pri-
vate households (90.8%). The majority were married
(55.4%). Most of them had their own children (85.6%) with
an average of two. Furthermore, the majority of the partic-
ipants had completed an apprenticeship (27.6%) or profes-
sional school (23.4%). For more details see Table 1.
Table 1. Sample characteristics (N= 195)
Variable
Gender n(%) Female 110 (56.4)
Male 85 (43.6)
Age M72.5
Mdn 71.4
SD 8.8
Range 60–94
Nationality n(%) Austrian 165 (84.6)
German 30 (15.4)
Place of
residence
n(%) Austria 161 (82.5)
Germany 31 (15.9)
Missing 3 (1.5)
Housing n(%) House 85 (43.6)
Apartment 91 (46.7)
Private room 1 (0.5)
Senior citizens res-
idence
18 (9.2)
Family
status
n(%) Single 10 (5.1)
Relationship 12 (6.2)
Married 108 (55.4)
Separated 2 (1.0)
Divorced 19 (9.7)
Widowed 44 (22.6)
Children n(%) No children 28 (14.4)
M2.0
Mdn 1.9
SD 1.4
Range 0–8
% alive 95.2
Adoptive
or foster
children
n(%) 9 (5.2)
Educational
background
n(%) Compulsory
school
23 (12.0)
Apprenticeship 53 (27.6)
Professional
school (without
Matura)
45 (23.4)
Secondary school
(with Matura)
29 (15.1)
College 8 (4.2)
University 33 (17.2)
A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults 33
GeroPsych 24 (1) © 2011 Hogrefe Publishing

Procedure
The empirical examination of the three newly developed
scales was part of a larger project, which also included
questionnaires on mental and physical health as well. Due
to issues associated with the older ages of the sample group
(e.g., some of them were visually impaired), 18 participants
(9.2%) answered the questions in the context of an inter-
view (items were read to them out loud). In these cases the
interview usually took place at the home of the participants.
However, most of the subjects (90.8%), completed the
questionnaires on their own.
After 1 month, part of the sample group (n= 20) that
had given their agreement when first interviewed were sent
the questionnaires on generativity again by mail to test the
stability of the construct; 19 of them returned them back
after, on average, 41.3 days (SD = 8.1, Min = 31/Max = 59
days). The selection criterion was that they were able to
complete the questionnaires on their own.
Development of Three New Scales (Gen-Life,
Gen-Current, and Gen-Parental)
The three new questionnaires are based on Kotre’s (1984)
four types of generativity (biological, parental, technical,
and cultural) complemented by social and ecological gen-
erativity. They focus on generative concerns in the life re-
view (Gen-life), current generative concerns (Gen-current)
and past parental generative concerns (Gen-parental). Fur-
thermore, these are self-assessment instruments in which
answers are given on a 4-point Likert-scale (from 1 = does
not apply to 4 = applies). Participants received the infor-
mation that the questionnaires were on generativity and
their own self-perception. After the scales had been pretest-
ed, only a few minor changes were made.
Generative Concerns in the Life Review
Gen-life consists of 29 items with four generativity do-
mains: technical, cultural, social, and ecological. Individ-
uals were to average their generative concerns over their
whole adult life (beginning at the age of 20). An example
was included in the introduction. Items always completed
a sentence that began with, “It has been a concern for me
. . .” (e.g., “. . . to pass on experiences to younger people.”).
Younger people were defined as subsequent generations
(15–20 years younger).
Current Generative Concerns
Gen-current consists of the same 29 items as Gen-life,
again with the four generativity domains: technical, cultur-
al, social, and ecological. The only aspect of the instruc-
tions that differed from the previous questionnaire is that
items were now to be rated as if they applied to the present
moment.
Past Parental Generative Concerns
Gen-parental was created for individuals who were parents
or who had been parents (in cases in which their child or
children already died). The scale consists of two domains:
biological and parental generativity in the narrower sense,
with altogether 19 items. Individuals were to indicate sev-
eral concerns they have had when their child/children grew
up (i.e., from birth until the age of 20).
Loyola Generativity Scale (LGS)
Completing the LGS, adults rated how often 20 statements
(e.g., “I try to pass along the knowledge I have gained
through my experiences.”) applied to them on a 4-point
scale ranging from never to very often/nearly always. Cron-
bach’s αs of .83 and .84, respectively, suggest a high inter-
nal consistency in two different samples (an adult sample
ranging in age from 19 to 68 and a college sample). Test-
retest reliability of the LGS over the 3-week period was
.73, suggesting moderately high temporal stability (Mc-
Adams & St. Aubin, 1992). The LGS contains two items
that are problematic for older adults, namely, “If I were
unable to have children of my own, I would like to adopt
children” and “I think I would like the work of a teacher.”
For this study the first item was deleted and the second one
changed into “I think I liked or would have liked the work
of a teacher.” Thus, the final version of the LGS used in
this study included 19 items.
Social Desirability
A total of 6 out of 17 items of the social desirability scale
previously used by Ochse and Plug (1986) in a study with
a self-reporting questionnaire of Erikson’s first seven stag-
es were included, e.g., “I am completely honest with ev-
erybody.” As social desirability was mixed in with the LGS
items, both scales had the same answering categories.
Results
Because preliminary investigations as well as more litera-
ture supporting the assumed factors would have been re-
quired to use confirmatory factor analyses, this study used
exploratory factor analysis (principal component analysis).
First, we used both oblique (oblimin) and orthogonal (vari-
max) factor analysis to examine the data. If both rotations
lead to similar results, a very stable factor structure could
34 A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults
GeroPsych 24 (1) © 2011 Hogrefe Publishing

Table 2. Gen-Life: median, loadings, corrected item-total correlations, explained variance and internal consistencies (n=
165)
Rotated factor analysis (Varimax)
Components
Item Mdn 1234rita
Social G.
29 Take on responsibility 3.1 .79 .15 .18 .01 .73
8 Care for young people 3.2 .74 .14 –.01 –.07 .61
19 Help young people to develop 3.4 .68 .07 .35 .13 .64
10 Guide young people 2.8 .67 .13 .10 .14 .59
7 Assist young people 3.1 .65 .14 .33 .12 .62
12 Impart social values 3.4 .55 .42 .27 .21
17 Pass on knowledge 3.3 .52 .24 .26 .25
15 Be a role model 3.3 .48 .33 .03 .13 .44
6 Impart cultural values 3.1 .48 .18 .45 .05
5 Teach how to play certain games 2.7 .46 –.08 –.08 .24
25 Show how to handle money 3.3 .39 .36 –.12 .26
1 Pass on experiences 3.4 .33 .24 .13 .25
Ecological G.
24 Be thrifty with energy 3.5 .00 .81 .08 .11 .64
18 Leave behind an intact environment 3.5 .19 .79 .15 .06 .68
14 Live ecology-minded 3.5 .17 .78 .14 .13 .68
3 Avoid garbage 3.5 .04 .64 –.08 .06 .49
26 Buy organic food 2.9 .15 .50 .19 –.07 .38
22 Protect animals 3.3 .18 .49 –.26 .34 .42
4 Support social institutions 3.1 .27 .45 .35 –.19
Cultural G.
20 Positive changes in society 2.7 .22 .21 .73 .23 .66
28 Pass on political values 2.6 .13 .12 .73 .16 .56
2 Get involved politically 2.0 –.15 –.01 .71 .00 .40
21 Support cultural facilities 2.4 .36 .02 .51 –.04 .43
9 Talk with young people about future of humankind 2.8 .35 .20 .44 .27 .45
16 Create enriching works 2.0 .34 –.13 .43 .26 .45
Technical G.
13 Teach how to repair certain things 2.8 –.01 .21 .08 .76 .63
27 Teach how to deal with technical devices 2.3 .10 –.05 .18 .75 .57
11 How to work with certain materials 2.8 .20 .06 .16 .73 .65
23 Pass on skills 3.3 .49 .17 .01 .63 .53
Explained variance % 17.0 13.2 11.1 9.8
(absolute) 33.3 25.9 21.7 19.2
Cronbach’s α.83 .79 .75 .78
Note. Response format: 1 = does not apply, 2 = rather does not apply, 3 = rather applies, 4 = applies. arit≥.20, p< .01.
A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults 35
GeroPsych 24 (1) © 2011 Hogrefe Publishing

be assumed and an orthogonal factor solution recommend-
ed for ease of interpretation (e.g., Bühner, 2004). Advanta-
ges of an orthogonal rotation include lower intercorrela-
tions between scales and greater independence of results
from observed samples. As intercorrelations of scales were
quite similar and relatively low in the current study for
oblique and orthogonal rotation, we chose varimax rotation
to maximize the independence of the scales for item anal-
ysis. It was not always possible to satisfy all the criteria,
e.g., some items had higher loadings on other factors; yet
some of these had to be kept in the scale as a minimum
number of items were needed in order to satisfy internal
consistencies as close as possible to the criteria Cronbach’s
α≥.80 (Bühner, 2004). Scale scores were built upon added
raw scores.
Generative Concerns in the Life Review
(Gen-Life)
For Gen-life, four factors were predicted: technical and
cultural generativity based on Kotre (1984) and Snarey
(1993) complemented by social and ecological generativ-
ity. Factor analysis yielded seven factors with initial ei-
genvalues over 1. Based on the scree plot showing a de-
cline after the fourth factor, and in accordance with the-
oretical assumptions (see above), the four-factor solution
was reported as follows: The eigenvalue of the first factor
was 8.4 (28.8% explained variance), the second factor 2.5
(8.8% explained variance), the third factor 2.0 (7.0% ex-
plained variance), and the fourth factor 1.8 (6.4% ex-
plained variance). Thus, explanation of variance for the
fourfactorsis51.0%.
For details concerning explained variance after rotation
see Table 2. Of the 12 items with the highest loadings on
factor 1 eight were created for social, two for cultural (6,
17), and two for technical (5, 25) generativity. The four
items not created for this scale will not be considered in
further analysis, nor will the two items (1, 12) created for
social generativity that display similarly high loadings on
other factors. Of the seven items with the highest loadings
on factor 2, six were developed for ecological generativ-
ity. Item four was created for social generativity and will
not be considered in further discussions of this scale. All
six items with the highest loadings on factor 3 were cre-
ated for cultural generativity and will be used for that
scale. The four items with the highest loadings on factor
4 were developed for technical generativity. Item 23 is
problematic because of a high loading on factor 1 but was
kept in the scale, as it only consists of four items.
Corrected item-total correlations of the four scales were
all highly significant (rit ≥.38, p< .01) and the internal
consistency was satisfying (Cronbach’s α≥.75). For more
details see Table 2. In sum, of the 29 Gen-life items, 22 can
be assigned to the presumed scales.
Current Generative Concerns (Gen-Current)
As for the scale on generativity in the life review, a four-
factor solution was predicted for current generativity. Fac-
tor analysis yielded five factors with initial eigenvalues
over one. According to the scree plot and theoretical as-
sumptions, the four-factor solution was reported in the fol-
lowing manner: The eigenvalue of the first factor was 8.9
(30.6% explained variance), the second factor 3.0 (10.3%
explained variance), the third factor 2.0 (7.0% explained
variance), and the fourth factor 1.6 (5.3% explained vari-
ance). Altogether, explanation of variance for the four fac-
tors is 53.2%.
For details concerning explained variance after rotation
see Table 3. Five of the six items with the highest loadings
on factor 1 were developed for technical generativity. Item
25 was developed for social generativity and will not be
considered in further analyses. Item 20 has a high loading
on factor 3 as well but will be included in the scale as the
distance between loadings is large enough. Seven of the 10
items with the highest loadings on factor 2 were created for
cultural generativity. Item 27 will not be included in further
discussions because of its high loading on factor 3. The
remaining three items (12, 14, 11) were developed for so-
cial generativity and will not be considered in further anal-
yses. Of the seven items with the highest loadings on factor
3, five were developed for social generativity; one was cre-
ated for cultural (18) and one for technical (8) generativity.
All items developed for social generativity will be consid-
ered in further analyses. Unfortunately, most of these have
high loadings on other factors as well (items 3, 6, 17, and
19). All six items with the highest loadings on factor 4 were
created for the ecological generativity scale. Although,
item 10 also has a high loading on factor 2, the distance
between the loadings is large enough thus it will be consid-
ered in further analyses.
Corrected item-total correlations of the four scales were
all highly significant (rit ≥.36, p< .01) and internal consis-
tency was satisfying (Cronbach’s α≥.74). For more details
see Table 3. Altogether, of the 29 items of the Gen-current,
again 22 were allocated to the presumed scales.
Past Parental Generative Concerns
(Gen-parental)
Based on Kotre’s theory (1984) a two-factor solution was
predicted for parental generativity (biological and paren-
tal). The scree plot supports a two- and three-factor solu-
tion. Based on theoretical assumptions, the two-factor so-
lution was reported in the following manner: The eigenval-
ue of the first factor was 5.8 (30.6% explained variance)
and the second factor 1.9 (9.8% explained variance).
Hence, explanation of variance for both factors together is
40.4%. For details concerning explained variance after ro-
tation see Table 4.
36 A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults
GeroPsych 24 (1) © 2011 Hogrefe Publishing

Table 3. Gen-current: median, loadings, corrected item-total correlations, explained variance, and internal consistencies
(n= 157)
Rotated factor analysis (Varimax)
Components
Item Mdn 1234rita
Technical G.
13 Teach how to repair certain things 2.6 .82 .07 .05 .13 .74
9 How to work with certain materials 2.6 .74 .12 .12 .13 .68
15 Teach how to deal with technical devices 2.2 .73 .03 .18 –.04 .67
20 Pass on skills 3.0 .65 .12 .49 .03 .64
25 Care for young people 3.1 .60 .27 .24 .10
23 Teach how to play certain games 2.5 .53 .22 –.03 .14 .36
Cultural G.
22 Pass on political values 2.7 .18 .69 .10 .08 .55
26 Get involved politically 1.9 .08 .64 –.06 –.10 .50
7 Support cultural facilities 2.5 .02 .59 .17 .01 .43
12 Take on responsibility 2.9 .49 .58 .21 .00
4 Positive changes in society 2.9 .05 .54 .28 .24 .47
14 Support social institutions 3.0 .05 .53 .16 .13
2 Create enriching works 1.8 .19 .53 .11 –.06 .44
27 Impart cultural values 3.0 .14 .52 .49 .09
29 Talk with young people about future of humankind 3.0 .25 .52 .25 .12 .45
11 Assist young people 3.2 .44 .48 .22 .07
Social G.
1 Pass on experiences 3.1 .10 .13 .75 .06 .55
3 Impart social values 3.3 .04 .39 .65 .23 .59
18 Pass on knowledge 3.3 .39 .27 .65 .13
6 Be a role model 3.5 .04 .06 .61 .34 .46
17 Help young people to develop 3.2 .47 .33 .58 .02 .68
19 Guide young people 2.7 .44 .35 .56 –.07 .62
8 Show how to handle money 3.1 .29 .07 .48 .27
Ecological G.
16 Live ecology-minded 3.6 .00 .09 .19 .86 .73
28 Avoid garbage 3.7 .08 –.02 .05 .83 .67
21 Be thrifty with energy 3.6 .13 .03 .03 .81 .60
5 Leave behind an intact environment 3.6 –.05 .22 .27 .67 .57
24 Protect animals 3.4 .18 –.09 .28 .56 .49
10 Buy organic food 2.9 .15 .35 –.07 .50 .36
Explained variance % 14.3 13.8 13.0 12.1
(absolute) 26.9 26.0 24.4 22.7
Cronbach’s α.82 .74 .80 .78
Note. Response format: 1 = does not apply, 2 = rather does not apply, 3 = rather applies, 4 = applies. arit ≥.20, p< .01.
A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults 37
GeroPsych 24 (1) © 2011 Hogrefe Publishing

All 10 items loading highest on factor 1 were developed
for the parental generativity scale. Item 10 will not be in-
cluded in the final version of the scale as it shows high
loadings on both factors. All four items created for the bi-
ological generativity scale have the highest loadings on fac-
tor 2 and will be considered in further analyses.
Corrected item-total correlations of the items included
were highly significant (rit ≥.28, p< .01). Cronbach’s αof
.68 for biological generativity was still acceptable taking
into account the fact that the scale consisted of only four
items. Internal consistency of α= .82 for parental genera-
tivity was satisfying.
Altogether, of the 19 items of the Gen-parental, 13 could
be assigned to the presumed scales.
Intercorrelations
Intercorrelations of Scales
Scales Within Scales
–Gen-life: As can be seen in Table 5, all correlations be-
tween the scales of the four-factor solution are signifi-
cant (range: .24 ≤r≤.49, all p< .01), suggesting low to
moderate degrees of association. The highest correlation
was found between social and cultural generativity (r=
.49, p< .01); this correlation of middle height demon-
strates that the two domains are still relatively independ-
ent from each other.
–Gen-current: As with the Gen-life questionnaire, all cor-
Table 4. Gen-parental: median, loadings, corrected item-total correlations, explained variance, and internal consistencies
(n= 156)
Rotated factor analysis (Varimax)
Components
Item Mdn 12rita
Parental G.
12 Support friendships 3.2 .74 .05 .64
15 Deal with achievement 3.5 .70 .12 .60
13 Pass on values 3.7 .67 .25 .60
14 Encourage partnerships 3.3 .67 .21 .58
9 Support leisure activities 3.6 .63 .02 .50
19 Be a role model to deal with work 3.7 .62 .19 .53
3 Read or tell stories 3.5 .61 .04 .44
7 Support career decision 3.6 .49 .16 .42
17 Mind to whom friendships 3.3 .46 .30 .42
10 Pass on experiences 3.6 .45 .41
Biological G.
8 Continuity of family through offsprings 3.3 .18 .77 .62
2 Similar character of child/children 3.0 .13 .76 .55
16 Similar profession 1.4 .02 .59
11 Look similar 1.8 .03 .59 .42
1 Be a role model concerning marriage 3.7 .17 .53
18 Procure family identity 3.5 .45 .46
4 Prepare for own family life 3.3 .41 .45
5 Have own biological child/children 3.8 .13 .45 .28
6 Care for relationships to relatives 3.4 .35 .40
Explained variance % 22.7 17.7
(absolute) (56.3) (43.8)
Cronbach’s α.82 .68
Note. Response format: 1 = does not apply, 2 = rather does not apply, 3 = rather applies, 4 = applies. arit ≥.21, p< .01.
38 A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults
GeroPsych 24 (1) © 2011 Hogrefe Publishing

relations between the scales of the four-factor solution
are significant (range: .25 ≤r≤.52, all p< .01). These
correlations again suggest only low to middle levels of
association between the scales. The highest correlations
could be found between social and technical (r= .52, p
< .01), and between social and cultural (r= .52, p< .01)
generativity. Much like the Gen-life results, these corre-
lations of middle height suggest that the different do-
mains are relatively independent from each other.
–Gen-parental: The correlation between parental and bi-
ological generativity of r= .39 (p< .01) suggests that
the two scales are associated with each other, but only
moderately.
Scales Between Scales
–Gen-life and Gen-current: For the four-factor solution,
as expected, the highest correlations (range: .53 ≤r≤
.80, all p< .01) were found between the same scales of
the two different questionnaires (e.g., technical scale
Gen-life and Gen-current). In particular, the associations
between the cultural (r= .80, p< .01) and ecological (r=
.77, p< .01) scales are high enough to question whether
or not substantial additional information is provided by
the two time perspectives measured (for details see Table
5).
–Gen-life and Gen-parental: Between the four scales of
the Gen-life and the two scales of the Gen-parental, cor-
relations are consistently stronger with parental (range:
.25 ≤r≤.39, all p< .01) versus biological generativity
(only the correlation between biological and social gen-
erativity is significant with r=.22,p< .01) as can be
seen in Table 5.
–Gen-current and Gen-parental: A correlation pattern,
similar to the scales of the Gen-life and Gen-parental,
appears between the scales of the Gen-current and Gen-
parental (see Table 5).
In sum, parental generativity seems to be more closely as-
sociated with other aspects of generativity than biological
generativity.
Correlations of Scales with the Loyola Generativity
Scale (LGS) and Social Desirability
LGS and New Scales
Correlations of low to middle height between the scales of
the three new questionnaires and the LGS (range: .20, p<
.05 ≤r≤.55, p< .01) indicate a positive convergent valid-
ity result (see Table 5). The scales are associated with each
other but not at such a high level as to suggest that they are
measuring the same aspects of generativity.
Social Desirability and New Scales
Findings indicate that parental and biological generativity
are possibly more strongly influenced by social desirability
(r= .29 and .33, both p< .01) than the other domains of
generativity (range: .04, p> .05 ≤r≤.23, p< .01). How-
ever, the LGS was the scale most strongly associated with
social desirability in the current study (r= .39, p< .01),
maybe because social desirability was mixed with the LGS
items. For more details see Table 5.
Table 5. Intercorrelations of scales of the three new scales, the LGS, and social desirability (N= 195)
Gen-life Gen-current Gen-parental LGS Soc. des.
(Sub)scales Techn. Cult. Social Ecolog. Techn. Cult. Social Ecolog. Parent. Biolog.
Gen-life
Techn. 1
Cult. .35** 1
Social .38** .49** 1
Ecolog. .25** .24** .40** 1
Gen-current
Techn. .71** .27** .24** .25** 1
Cult. .21** .80** .33** .18* .37** 1
Social .40** .43** .53** .25** .52** .52** 1
Ecolog. .27** .20** . 23** .77** .26** .25** .34** 1
Gen-parental
Parent. .25** .32** .39** .33** .25** .31** .39** .38** 1
Biolog. .12 .04 .22** .09 .16 –.02 .26** .12 .39** 1
LGS .32** .47** .44** .24** .42** .52** .55** .26** .34** .20* 1
Soc. Desirability .09 .04 .16* .18* .14 .12 .21** .23** .29** .33** .39** 1
Note.*p< .05, **p< .01.
A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults 39
GeroPsych 24 (1) © 2011 Hogrefe Publishing

Influence of Gender and Age (Two-Factor
ANOVA)
Gen-Life
The ANOVA yielded significant main effects for gender
concerning cultural generativity (F(1, 182) = 11.42, p<
.00; η2= .06), and technical generativity (F(1, 183) =
18.90, p< .00; η2= .10) in the life review, with men scoring
higher than women in all cases. No main effect for age
could be found. Furthermore, the ANOVA yielded one sig-
nificant interaction effect for cultural concerns in the life
review (F(1, 182) = 5.87, p= .02; η2= .03), with younger
men and women (60–74 years) having similar concerns
which diverged as they got older (older men having higher,
and older women lower scores, compared with the younger
age group).
Gen-Current
The ANOVA yielded significant main effects for gender
concerning current cultural generativity (F(1, 168) = 5.08,
p=.03;η2= .03), technical generativity (F(1, 177) =
18.17, p<.00;η2= .10), and social generativity
(F(1, 175) = 5.14, p= .03; η2= .03) with men scoring high-
er than women in all cases. No main effect for age and no
interaction effect could be found.
Gen-Parental
The ANOVA yielded no significant main effect concerning
gender. For age (age groups: 60–74 years and 75–94 years)
one significant main effect could be found concerning bi-
ological generativity (F(1, 163) = 8.45, p< .00; η2= .05),
with the older age group scoring higher. No interaction ef-
fect could be found for this scale concerning age and gen-
der.
LGS
In terms of a control analysis, a significant main effect con-
cerning gender could be found (F(1, 169) = 9.41, p< .00;
η2= .05) for the LGS, again with men scoring higher than
women. However, the ANOVA yielded no age or interac-
tion effect.
In sum, gender had a much stronger influence on gener-
ativity than age and only one interaction effect could be
found for these two variables.
Bivariate Correlations Between Age and
Generativity
Most correlations (Spearman) between scales on generativ-
ity and age were not significant (range: .01 ≤r≤–.15).
Only biological generativity was positively associated with
age, r= .33 (p< .01).
Difference Between Parents and Childless
Older Adults
The Mann-Whitney Test yielded only one significant dif-
ference concerning parenthood and generativity, with par-
ents (n1,M= 19.25, SD = 3.65) scoring higher than child-
less older adults (n2,M= 17.27, SD = 3.96) on social gen-
erativity (life review), U(n1= 153, n2= 26) = 1397.5, p<
.05.
Retest New Scales and LGS
Gen-Life
Even when taking into account the Spearman-Brown
prophecy (19 items as the LGS), only technical generativity
(r= .85, p< .01) showed very satisfying retest stability (see
Table 6). Ecological generativity practically showed no sta-
bility (r= .12, p> .05) between the two time points. In sum,
these results seem unconvincing with regards to the stabil-
ity of life review generativity.
Gen-Current
Spearman-Brown corrected test-retest correlations for cur-
rent generative concerns suggest a satisfying to very satis-
fying stability (range: .77 ≤r≤.95, all p< .01).
Table 6. Test (N= 195)-retest (N= 19)-correlations of the scales, the LGS, and social desirability
Gen-life Gen-current Gen-parental LGS
Techn. .54*a(.85b) .72** (.91) Parent. .62* (.77) .81**
Cult. .35 (.63) .73** (.90) Biolog. .60* (.88)
Social .28 (.55) .82** (.95)
Ecolog. .04 (.12) .52* (.77)
Notes.*p< .05, **p< .01. aSpearman-correlations, bSpearman-Brown prophecy (19 items).
40 A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults
GeroPsych 24 (1) © 2011 Hogrefe Publishing

Gen-Parental
For biological (r= .77, p< .01) and parental (r= .88, p<
.01) generativity, Spearman-Brown corrected correlations
between the first and second assessment were significant,
suggesting satisfying to very satisfying stability for these
scales.
LGS
For the LGS, a satisfying stability was found (r= .81, p<
.01).
Thus, except for some scales on generativity in the life re-
view (especially ecological generativity), all correlations
of the scales between the two measurement points show
satisfying to very satisfying stability.
Discussion
Although already identified as a topic associated with suc-
cessful aging (e.g., Erikson & Erikson, 1997; Fisher, 1995),
generativity has not received much attention. Most studies
have not involved adults beyond their early 70s, and there
is a lack of adequate instruments able to capture the unique
features associated with generativity in older age. There-
fore, new scales based on Kotre’s (1984) four age-specific
types of generativity were developed and empirically tested
with a sample of older adults.
Results give empirical evidence for technical, cultural,
social, and ecological generativity for the scales addressing
current generativity and generativity in the life review. In
regards to parental generativity, empirical support was
found for the two scales addressing biological and parental
generativity. Cronbach’s αas a criterion of internal consis-
tency are satisfying. Intercorrelations demonstrated low to
middle degrees of correlation for the scales within the three
newly developed questionnaires and support the assess-
ment of generativity in a multidimensional way. Despite
the small sample size of the second measuring point, all of
the scales – except for generativity in the life review –
showed strong stability over an average of 41.3 days. One
explanation could be that life review is not a stable con-
struct and more subject to influence (e.g., by current mood).
As results for parental generativity (concerning the past)
were higher, another reason could be that the instructions
given for the life review scale were ambiguous and caused
instability, as subjects were asked to average their concerns
over the whole adult life span. Future research should
therefore focus on exploring in greater detail how genera-
tivity in the life review could be assessed more reliably.
Given that the correlations between the same scales of
generativity in the life review and current generativity were
high enough to question whether or not the introduction of
both scales provides additional information, and the life
review scale showed a relatively weak stability, one con-
clusion of this study would be to use Gen-current rather
than Gen-life in future studies.
In direct contrast to the literature (e.g., McAdams & St.
Aubin, 1992; Peterson, Smirles, & Wentworth, 1997), men
consistently reported higher generativity scores in the cur-
rent study. One possible explanation for this is that predom-
inantly midlife adults were included in previous research.
Similar, Lugmayr (2010) found that grandchildren rated
their grandfathers as being more generative than their
grandmothers concerning parental, social, cultural, and
technical generativity (measured with an adapted version
of the scales used in the current study). According to the
convergence hypothesis (Carmel & Bernstein, 2003), gen-
der differences should decline during aging because social
roles between the two genders become more similar and
parenthood and occupation lose their importance. It is pos-
sible that men are becoming more aware of their generative
concerns as they age, while women, on the other hand, may
focus on the implicitness of social concerns over the entire
lifespan. Support for this suggestion comes from Höpflin-
ger, Hummel, and Hugentobler (2006), who found that
child care assistance is regarded to be more of a matter of
course in grandmothers than in grandfathers. Consequent-
ly, perception is stronger when grandfathers are involved
with their grandchildren. Another explanation for the fact
that men scored higher than women could be that some
aspects are potentially more stereotypically male (especial-
ly technical and cultural generativity). However, gender
differences were found for social generativity as well.
Thus, the role that gender plays in the aging process needs
to be further investigated, perhaps by investigating various
age groups and the link between changing gender roles re-
spectively gender role flexibility and their relationship to
generativity.
Older subjects in the sample (75 years and older) report-
ed having had more biological concerns. One explanation
for this could be the fact that these individuals were young
adults after WWII, a period that certainly involved some
childbearing pressure.
Associations between the three newly developed scales
and the LGS, the most common instrument measuring gen-
erativity demonstrate a positive validation result. However,
the scales are not correlated with the LGS at a high enough
level to suggest that they are measuring the same aspects
of generativity. These results support the value of introduc-
ing these new scales in order to complement the most fre-
quently used instrument to measure generativity.
Social desirability had particularly strong associations
with social, ecological, parental, biological generativity,
and the LGS. Findings of the current study clearly suggest
that social desirability must be controlled for when inves-
tigating generativity.
Because the results are, in general, based on self-report-
ing, it is not possible to estimate to what extent participants
really are currently generative, nor the extent to which they
A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults 41
GeroPsych 24 (1) © 2011 Hogrefe Publishing

have been in the past. Making use of other-rating systems
(e.g., by asking the subject’s children) might be one way in
which self-reporting data could be regarded as more valid.
However, it seems crucial to consider and incorporate the
self-concept of older adults when looking at aspects asso-
ciated with successful aging, particularly given the fact that
previous studies consistently found the strongest associa-
tions between generative concerns (and not e.g., actions)
and well-being (e.g., McAdams et al., 1993).
It can be concluded that these three new scales, which
were developed and empirically tested in order to assess
generativity among older adults, are important resources.
As hypothesized, age-specific scales (Kotre, 1984) and
their association with the most common instrument used to
measure generativity (LGS) was confirmed. Thus, these
new scales provide researchers with the unique opportunity
to measure generativity in older adults in a multidimension-
al way.
Acknowledgments
We are grateful to Erika Hütter-Schrenk, Christine Neuner,
Sabine Lüdtke, and Edith Wührer for their help in data col-
lection and data entry. We would like to thank Verena
Graupmann and Michelle Wright for their helpful com-
ments on an earlier version of this manuscript.
References
Ackerman, S., Zuroff, D.C., & Moskowitz, D. S. (2000). Gener-
ativity in midlife and young adults: Links to agency, commu-
nion, and subjective well-being. International Journal of Ag-
ing and Human Development, 50, 17–41.
An, J. S., & Cooney, T. M. (2006). Psychological well-being in
mid to late life: The role of generativity development and par-
ent-child relationships across the lifespan. International Jour-
nal of Behavioral Development, 30, 410–421.
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.), Success-
ful aging: Perspectives from the behavioral sciences
(pp. 1–34). Cambridge: Cambridge Univ. Press.
Bühner, M. (2004). Einführung in die Test- und Fragebogenkon-
struktion [Introduction to testing and scale construction]. Mün-
chen: Pearson.
Carmel, S., & Bernstein, J. (2003). Gender differencesin physical
health and psychosocial well-being among four age-groups of
elderly people in Israel. International Journal of Aging and
Human Development, 56, 113–131.
Dittmann-Kohli, F. (1995). Das persönliche Sinnsystem [The per-
sonal meaning system]. Göttingen: Hogrefe.
Erikson, E. H. (1950). Childhood and society. New York: Norton,
Inc.
Erikson, E. H., & Erikson, J. M. (1997). The life cycle completed:
Extended version with new chapters on the ninth stage of de-
velopment. New York: Norton & Comp.
Erikson, J. M., Erikson, E. H., & Kivnick, H. (1986). Vital involve-
ment in old age. New York: Norton & Comp.
Fisher, B.J. (1995). Successful aging, life satisfaction, and gen-
erativity in later life. International Journal of Aging and Hu-
man Development, 41, 239–250.
Hagestad, G. O., & Uhlenberg, P. (2006). Should we be concerned
about age segregation? Some theoretical and empirical explo-
rations. Research on Aging, 28, 638–653.
Hoare, C. H. (2002). Erikson on Development in Adulthood: New
Insights from the Unpublished Papers. New York: Oxford Uni-
versity Press.
Höpflinger, F., Hummel, C., & Hugentobler, V. (2006). Enkelkin-
der und ihre Großeltern: Intergenerationelle Beziehungen im
Wandel [Grandchildren and their grandparents: Changing gen-
erational relationships]. Zürich: Seismo.
Keyes, C. L. M., & Ryff, C. D. (1998). Generativity and adult
lives: Social structural contours and quality of life consequenc-
es. In D. P. McAdams & E. de St. Aubin (Eds.), Generativity
and adult development: How and why we care for the next
generation (pp. 227–263). Washington, DC: American Psy-
chological Association.
Kotre, J. (1984). Outliving the self: Generativity and the interpre-
tation of lives. Baltimore: Johns Hopkins University Press.
Kotre, J. (1996). Outliving the self: How we live on in future gen-
erations. New York: Norton & Co.
Kruse, A., & Wahl, H.-W (2009). Zukunft Altern: Individuelle und
gesellschaftliche Weichenstellungen [The future of aging: In-
dividual and societal directions]. Heidelberg: Spektrum.
Lang, F., & Baltes, M. (1997). Brauchen alte Menschen junge
Menschen? Überlegungen zu den Entwicklungsaufgaben im
hohen Lebensalter. In L. Krappmann & A. Lepenies (Hrsg.),
Alt und jung. Spannung und Solidarität zwischen den Genera-
tionen [Old and young: Tension and solidarity between gen-
erations] (pp. 161–184). Frankfurt/M.: Campus Verlag.
Lugmayr, F. (2010). Großelternschaft und großelterliche Gene-
rativität aus der Sicht der Enkel und Enkelinnen [Grandparen-
thood and generativity from the point of view of grandchil-
dren]. Unveröffentlichte Diplomarbeit, Universität Salzburg.
Manheimer, R.J. (1995). Redeeming the aging self: John Kotre,
George Drury, and Cultural Generativity. Journal of Aging
Studies, 9, 13–20.
McAdams, D., & St. Aubin, E. de (1992). A theory of generativity
and its assessment through self-report, behavioral acts, and
narrative themes in autobiography. Journal of Personality and
Social Psychology, 62, 1003–1015.
McAdams, D., St. Aubin, E. de, & Logan, R. (1993). Generativity
among young, midlife, and older adults. Psychology and Ag-
ing, 8, 221–230.
McKeering, H., & Pakenham, K. I. (2000). Gender and genera-
tivity issues in parenting: Do fathers benefit more than mothers
from involvement in child care activities? Sex Roles, 43,
459–480.
Ochse, R., & Plug, C. (1986). Cross-cultural investigation of the
validity of Erikson’s theory of personality development. Jour-
nal of Personality and Social Psychology, 50, 1240–1252.
Peterson, B.E., Smirles, K. A., & Wentworth, P.A. (1997). Gen-
erativity and authoritarianism: Implications for personality,
political involvement, and parenting. Journal of Personality
and Social Psychology, 72, 1202–1216.
Snarey, J. (1993). How fathers care for the next generation. Cam-
bridge: Harvard Univ. Press.
42 A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults
GeroPsych 24 (1) © 2011 Hogrefe Publishing

St. Aubin, E. de, & McAdams, D. (1995). The relations of gen-
erative concern and generative action to personality traits, sat-
isfaction/happiness with life, and ego development. Journal of
Adult Development, 2, 99–112.
Vaillant, G. E. (2002). Aging well: Surprising guideposts to a hap-
pier life from the landmark Harvard study of adult develop-
ment. Boston: Little, Brown & Co.
Mag. Angela Schoklitsch
Department of Psychology
University of Salzburg
Hellbrunnerstraße 34
AT-5020 Salzburg
Austria
E-mail: angela.schoklitsch@stud.sbg.ac.at
A. Schoklitsch & U. Baumann: Measuring Generativity in Older Adults 43
GeroPsych 24 (1) © 2011 Hogrefe Publishing