Journal of Personality and Social Psychology Copyright 1996 by the American Psychological Association, Inc.

1996. Vol.70. No. 5. 1067-1079 0022-3514/96/53.00

Conceptual Analysis and Measurement of the Construct

of Ego-Resiliency
Eva C. Klohnen
University of California, Berkeley

This investigation explored the meaning, measurement, and validity of the ego-resiliency construct
(ER,J. H. Block &J, Block) in 3 samples. Study I explored the internal structure ofER in observer
and self-report data, and the development of a self-report measure is described. Study 2 tested con-
vergent and discriminant relations of ER with personality attributes. Study 3 investigated implica-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

tions of ER for adjustment and eifective functioning in adulthood. Components of ER obtained

This document is copyrighted by the American Psychological Association or one of its allied publishers.

through exploratory factor analyses—confident optimism, productive and autonomous activity, in-
terpersonal warmth and insight, and skilled expressiveness—formed a unitary construct and mir-
rored the relations found between ER and other trait domains. Across samples, there were strong
relations between ER and effective functioning in diverse areas of life. ER is discussed in relation to
generally accepted criteria of adjustment and eifective functioning.

Although every individual has to deal with minor hassles and identify the components of ER, (c) establish ER's convergent
major stressors, negotiate life transitions, and adapt to life and discriminant relations with other personality characteris-
changes, some individuals are much more competent than oth- tics, and (d) investigate the implications of ER for adjustment
ers in effectively coping with difficult times. This fact of life has and well-being in two samples of adult women.
provided an incentive for theorists and researchers to explore
the nature of these individual differences in adaptability in Definition, Operationalization, and Measurement of ER
hopes of ultimately being able to foster resilience and to prevent
the development of vulnerabilities. In a major review article, J. H. Block and Block (1980) elab-
There are many different approaches to the conceptualiza- orated the meaning they had assigned to the construct of ER
tion and operational ization of factors underlying human adapt- and its relations to other concepts. On the broadest level, ER
ability. Examples of related constructs include ego-strength, refers to the general capacity forflexibleand resourceful adap-
competence, emotional stability, coping or defense, self-efficacy, tation to external and internal stressors. More specifically, ER is
hardiness, and social intelligence. The present research focused a personality resource that allows individuals to modify their
on the construct of ego-resiliency (ER) as formulated by Jack characteristic level and habitual mode of expression of ego-con-
and Jeanne Block (J. Block, 1950; J.H. Block, 1951;J.H. Block trol so as to most adaptively encounter, function in, and shape
& Block, 1980). ER was initially conceptualized in the context their immediate and long-term environmental contexts. J. H.
of personality development and is a conceptually and theoreti- Block and Block (1980) denned ER as
cally well-grounded construct that accounts for dynamic per-
sonality processes. This investigation focused on the meaning, resourceful adaptation to changing circumstances and environ-
measurement, and construct validity of ER. More specifically, I mental contingencies, analysis of the "goodness offit"between sit-
(a) describe Block's Q-sort-based operationalization of ER as uational demands and behavioral possibility, and flexible invoca-
tion of the available repertoire of problem-solving strategies
well as an empirically constructed self-report measure, (b) (problem-solving being defined to include the social and personal
domains as well as the cognitive). The opposite end of the ego-
resiliency continuum (ego-brittleness) implies little adaptive flex-
ibility, an inability to respond to the dynamic requirements of the
This research was facilitated by a Jacob K. Javits Predoctoral Fellow- situation, a tendency to perseverate or to become disorganized
ship and a grant from the Spring Foundation. Further support was pro- when encountering changed circumstances or when under stress,
vided by Grant MH43948 from the National Institutes of Health. and a difficulty in recouping after traumatic experiences, (p. 48)
I am indebted to Ravenna Helson, Abigail Stewart, and Harrison
Gough and the Institute of Personality and Social Research for provid- The main operationalization of the ER construct has been an
ing me access to their data sets. I thank Ravenna Helson and Oliver John observer measure based on the California Adult Q-Set (CAQ; J.
for their advice and comments. I also profited from comments on an Block, 1961/1978). In this procedure, observers or raters who
earlier version provided by Jack Block, Gail Agronick, Sam Gosling,
Jennifer Pals, and Rick Robins. have familiarized themselves with the participants (either
Correspondence concerning this article should be addressed to Eva through direct interpersonal contact or through reading of ar-
C. Klohnen, University of California, Institute of Personality and So- chival materials collected on or provided by the participants)
cial Research, 2150 Kittredge Street, #5050, Berkeley, California provide personality descriptions of the participants via the Q-
94720-5050. Electronic mail may be sent via the Internet to sort procedure using the set of 100 Q-sort items. Typically, de-
kruemel@cmsa.berkeley.edu. scriptions of multiple raters are averaged to derive a more reli-
1067
 1068 KLOHNEN

able Q composite for each participant. As a means of establish- is likely to shed some light on the conceptual relations of ER
ing ER scores for the participants, the degree of congruence be- to other important constructs that tap into aspects of positive
tween participants' Q composites and a theoretically derived health, adaptability, and effective functioning.
prototype (or criterion) definition of ER is calculated. Appen-
dix A provides a listing of the 26 most denning CAQ items of Relations of E R to Specific Personality Attributes
the ER criterion definition (J. Block, 1991), including the 13
most characteristic and the 13 most uncharacteristic items of A related question concerns the relations of ER to other, more
the ER prototype.' specific personality domains or attributes. Reviews of the literature
Observer-based CAQ measures have proven to be valid and on resilience (e.g., Garmezy, 1985; Werner & Smith, 1982) con-
useful in many ways and in many contexts (e.g., J. Block, 1971; sistently point to a number of individual-differences variables that
Caspi & Herbener, 1990; Funder & Colvin, 1988; Ozer & appear to be among the core protective factors for children in
Gjerde, 1989; York & John, 1992). However, Q sorting is a stressful life situations. The temperamental and personality attri-
time-consuming procedure, and ratings by multiple observers butes consistently found to be associated with stress resistance in
are required. As a result, CAQ descriptions are not feasible in children are reflectiveness when confronted with new situations,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

most investigations. Thus, a self-report measure that proves to cognitive skills, activity level, and positive responsiveness to others.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

be comparable to the CAQ measure of ER would be very useful. These attributes are consistent with some of the behavioral im-
One of the aims of this article is to describe the development plications the Blocks associated with ER, such as resourcefulness
and psychometric properties of such a self-report scale. This and integrated performance under stress, adaptive flexibility, ac-
scale was derived empirically by means of criterion keying, with tive engagement with the world, and an available repertoire of
the observer-based CAQ ER measure serving as the criterion problem-solving strategies within the social, personal, and cogni-
and the California Psychological Inventory (CPI; Gough, 1987) tive domains. Accordingly, one would expect ER to encompass
providing the pool from which the items were drawn. The CPI (and to be related to) aspects of personality that tap into adjust-
item pool is well suited for the purpose of constructing an ER ment, effective coping and intellectual functioning, flexibility, and
scale because the inventory is designed to measure effective psy- an engaging and active approach to the world. However, in line
chological and interpersonal functioning, aspects inherently im- with J. H. Block and Block's (1980) conceptualization of ER, as-
portant to the notion of ER. In addition, the CPI is a popular pects of personality indicative of overcontrol or undercontrol of
and widely used instrument, thus making the scale available in various needs and impulses should be independent of (i.e., unre-
a broad range of research contexts. lated to) ER because neither extreme provides a consistently adap-
tive approach.
I first analyzed the internal structure and major components of
Components of E R ER in an exploratory factor analysis of both the established CAQ
Over the last decade, the term resilience has come into in- observer measure and the new self-report scale. In a second step, I
creasing use, most notably by Garmezy (e.g., Garmezy, 1985; used confirmatory factor analysis to determine whether ER can be
Masten,Best,&Garmezy, 1990) and Rutter(e.g., 1979, 1987), considered a unitary construct. Finally, I tested the Blocks' theo-
who have focused on "stress-resistant" children. These re- rizing about the construct by establishing the convergent and dis-
searchers use resilience as a descriptive label that they apply criminant relations of ER to selected scales of the Adjective Check
to individuals who appear to function surprisingly well under List (ACL; Gough &Heilbrun, 1983).
environmental conditions judged to be adverse and stressful.
Earlier research tended to focus on understanding the major Implications and Life Outcomes of E R
protective factors that positively influence or contribute to
ER has clear implications for individuals' adaptive capabilities
stress resistance or resilience. More recently, attempts have been
under conditions of environmental stress, conflict, or uncertainty.
made to understand the protective processes underlying the
Consequently, resilience has primarily been examined within the
phenomenon of resilience (e.g., Egeland, Carlson, & Sroufe,
context of coping with extreme adversity (e.g., Compas, 1987;
1993;Rutter, 1987).
Cowen, Wyman, Work, & Parker, 1990; Garmezy, 1991; Rutter,
Overall, however, little attention has been focused on clarify- 1979). However, ER, as conceptualized by the Blocks, can also be
ing the conceptual meaning and the components of stress resis- expected to play a role in individuals' everyday experiences and
tance and resilience when these constructs are conceptualized their dealings with the joys, challenges, and hardships of life. Work
as individual-differences variables. Although ER has been con- involving ER has been conducted in a broad range of research
ceptualized as an individual-differences variable, little is known contexts, and powerful and important effects of the construct have
about its underlying components as well. An exploration of the emerged. For example, ER has been found to be related to secure
internal structure of ER can advance understanding of why and attachment styles across time (Kobak & Sceery, 1988; Sroufe,
how some individuals successfully adapt to challenging or Carlson, & Shulman, 1993), to moderate delay of gratification
threatening circumstances, whereas others succumb. Examin- (Funder & Block, 1989), to influence judgability of personality
ing the intrapsychic resources available to the resilient individ- (Colvin, 1993), and to predict personality consistency over time
ual and those lacking in the ego-brittle person can provide in- (Asendorpf & Van Aken, 1991). However, this research has fo-
sight into the proximal psychological processes involved in
effective and competent mastering of life's challenges and hard-
ships. In addition, although the present study cannot speak to 1
The item placements for the 100 Q-sort items for the entire ER
this issue directly, a detailed analysis of the components of ER prototype are available from the author.
 CONSTRUCT OF EGO-RESILIENCY 1069
cused almost exclusively on children and adolescents. Little is were San Francisco Bay Area residents (M age = 35.0 years, SD = 10.2).
known about the implications of ER for adaptive functioning in The others were 198 (99 of each sex) University of California at Berke-
adulthood. Therefore, this study tested the importance of ER for ley sophomores, most of whom were between 18 and 21 years of age
adjustment and effective functioning across diverse areas of life (mode = 1 9 ) . Q-set ratings and CPI data were available for all partici-
by examining its relations to a comprehensive, life-outcome-based pants in this sample.
Cross- Validation Sample I: Mills sample. This sample consisted of
measure of adjustment (Picano, 1989) in two samples of adult
the participants in a longitudinal study of Mills College graduates. A
women. representative two thirds of the 1958 and 1960 Mills graduating classes
(N = 141) were sampled. Helson (1967; Helson, Mitchell, & Moane,
Overview 1984; Helson & Wink, 1992) followed up these women on three differ-
ent occasions when the women were approximately 27,43, and 52 years
To recapitulate, the present investigation had four main ob- of age. The data examined in this study came from the age 43 assess-
jectives. Study 1, which involved the development of an ER self- ment, when Q-set ratings and CPI data were obtained for 105 and 107
report scale, examined the psychometric properties of the ER women, respectively.
scale and explored the components and structure of ER in both Cross-Validation Sample 2: Radcliffe sample. This sample con-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

the observer and the self-report measures. Both measures were sisted of the Radcliffe College graduating class of 1964. Stewart (1978,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

then used in Studies 2 and 3, which examined the validity of the 1980; Stewart & Salt, 1981; Stewart & Vandewater, 1993) has been fol-
construct. Specifically, Study 2 focused on the convergent and lowing the life trajectories of these women with five major data collec-
tions at ages 18,31,37,43, and 48. 2 The Radcliffe data examined in this
discriminant relations of ER to personality attributes, testing
study came from the 1986 (age 43; N = 103) and 1991 (age 48; N =
predictions by J. H. Block and Block's (1980) theory. Study 3 131) assessments.
tested the external criterion validity of the ER measures with
regard to adjustment and effective functioning as indexed by
life outcome data. 1 was concerned about generalizability issues; CAQ Personality Descriptions
thus, I used two measures of ER and replicated all findings The CAQ (J. Block, 1961 /1978) is an ipsative measure consisting of
across at least two samples. 100 descriptive statements that sample a broad domain of personality,
cognitive, and interpersonal characteristics and functioning. Each of the
statements is sorted by a rater into a fixed, quasi-normal, nine-category
Study 1: Internal Structure of ER in an Observer distribution ranging from extremely uncharacteristic to extremely char-
Measure and in a New Self-Report Scale acteristic in terms of its salience to the person being described.
Two hundred eighty members of the derivation sample had been de-
The purpose of thefirststudy was twofold: (a) to construct a scribed, through the use of the CAQ, byfiveto seven staff members who
self-report ER scale that approximates the Blocks' observer- had observed and interacted with them during weekend assessments at
based measure of ER and (b) to explore the internal structure IPSR. The remaining 70 participants had been interviewed by pairs
of both the observer and self-report instruments. As a means of of psychologists who provided the CAQ ratings. The interviews were
achieving the best possible self-report representation of the semistructured and included detailed questions about life history, cru-
CAQ ER measure, the development of the scale used observer cial life events, and personal and social values.
ER scores as the criterion. This approach avoided the difficult In the Mills and Radcliffe samples, a panel of three or more trained
task of having to establish, a priori, the possible specific mani- raters provided Q-sort descriptions for each participant based on the life
festations of ER in self-reports. Empirical keying was used in a data provided by 105 Mills and 103 Radcliffe women. The set of case
materials used as the basis for the ratings consisted of both open-ended
large, heterogeneous derivation sample to select items for the
questions and rating scales that elicited information about diverse areas
new scale from the CPI item pool. The findings on the con- of the women's lives, such as their parents, children, partners, careers,
vergent validity of the self-report with the observer criterion in goals, social roles, major life events, mental and physical health, and
the two replication samples are crucial. In addition, it is of cen- leisure activities. The mean alpha reliabilities for the composite CAQ
tral importance to show evidence for the incremental validity of ratings were .75 for the Mills sample and .76 for the Radcliffe sample
the ER self-report scale beyond existing CPI scales in predicting (for a more detailed discussion of the comparability of the Mills and
observer-based ER. Radcliffe samples, see Peterson & Klohnen, 1995).
The second part of Study 1 explored the components and the
internal structure of ER via exploratory factor analysis in the CAQ ER Criterion Scores
observer and the self-report measures. To test the present con-
As described by J. Block (1961 /1978, pp. 89-115), the CAQ, with
ceptualization of ER as a unitary yet broad construct that com- its wide range of content, can be used by experts to describe prototypes
bines several more specific components, I used confirmatory of various psychological constructs. Block (1991) had nine experts use
factor analysis to examine the viability of several competing the CAQ to describe a prototypical ego-resilient individual, thus speci-
models of the internal structure of the construct. fying the personality attributes thought to be associated with the con-
struct of ER. The criterion definers achieved high levels of agreement (a
= ,97), and the ratings were aggregated into a composite ER criterion
Method definition. The existence of the prototypical Q sort makes it possible to
Participants assess the level of ER of each participant by calculating the similarity

Derivation sample. The derivation sample consisted of 350 individ-

2
uals (175 men and 175 women) who had taken part in studies of per- Several waves of the data for this sample have been archived at the
sonality assessment at the Institute of Personality and Social Research Murray Research Center, Radclifle College, 10 Garden Street, Cam-
(IPSR)between 1972 and 1984. Of these participants, 152 (76 couples) bridge, Massachusetts 02138.
 1070 KLOHNEN

between the participant's Q description and the composite criterion de- to which the scale predicts the observer ER prototype scores in the
scription. A high correlation means that the individual is similar to the two independent replication samples. The validity coefficients for
prototypical definition (i.e., has high levels of ER); a low correlation the two samples were virtually identical: .62 in the Mills sample
means that the individual is dissimilar to the prototypical definition
and .59 in the Radcliffe sample. This is particularly impressive
(i.e., has low levels of ER). The similarity scores thus derived for each
given that, in the Radcliffe sample, the observer and self-report
individual can be taken as measures of his or her CAQ-defined level of
ER. This measure of "observer"-based ER criterion scores was used as measures were obtained at two time periods separated by 5 years.
the basis on which to select CPI items to be included in the self-report These coefficients are substantial replicated criterion validities, es-
measure of ER. pecially when compared with the coefficients in the .40s typically
observed in the literature on agreement between self-reports and
reports by others (for a recent review, see John & Robins, 1993).
CPI The magnitude of the obtained validities suggests that the ER scale
The CPI (Gough, 1957, 1987) is a 472-item self-report inventory can predict a substantial part of the reliable variance in the ob-
with 23 scales that provide a comprehensive set of personality measures; server ER measure.
themes of interpersonal adequacy, normative control of impulses, and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Even though the convergent validities just reported are sub-

achievement potential are included. The fact that the CPI is primarily a
This document is copyrighted by the American Psychological Association or one of its allied publishers.

stantial, it is possible that standard CPI scales might do equally

measure of effective interpersonal and psychological functioning ren-
well at predicting observer-based ER. I therefore tested (a)
ders the item pool particularly relevant for the construction of a mea-
whether the new ER scale is indeed the best predictor of ob-
sure of ER. CPI scores were available in all three samples (JVS = 350,
106, and 127 for the derivation, Mills, and Radcliffe samples, server-based ER and (b) whether it shows incremental validity
respectively). beyond the existing CPI scales across all samples. First, to test
whether the ER scale is the best predictor, I performed stepwise
multiple regressions predicting observer-based ER in all three
Development ofthe CPi-Based Self-Report Measure samples from the standard 23 CPI scales and from the ER scale.
ofER For all three samples, ER was indeed the strongest predictor;
moreover, none of the standard 23 CPI scales entered as signifi-
The derivation sample of 350 participants was used for initial selec-
tion and refinement of a set of CPI items to be included in the ER scale. cant predictors in all three samples. For example, the next
First, the array of 472 CPI items was correlated with the CAQ-based strongest predictors after ER were Self-Acceptance in the IPSR
ER criterion score for each participant. This analysis was performed (derivation) sample, Empathy in the Mills sample, and Com-
separately for the male and female participants. 3 A preliminary ER munality in the Radcliffe sample.
scale was constructed by including items that met the following criteria: Second, I examined whether self-reported ER showed incre-
They correlated at least .20 with CAQ ER within one sex and at least
mental validity over the standard CPI scales. In particular, 1
.15 within the other sex. These criteria resulted in an initial pool of
computed partial correlation coefficients between the ER scale
48 items. This preliminary item pool was further refined by internal
consistency analyses. Next, the item pool was correlated with the total and observer-based ER by controlling for the effects of those
scores on the provisional CPI ER scale and the standard CPI scales. To CPI scales that had entered as significant predictors in the step-
minimize unwanted overlap with other scales (i.e., to increase discrim- wise regressions just described. In all three samples, the ER
inant validity), I eliminated several items that were more highly corre- scale significantly {all ps < .001) predicted observer-based ER
lated with CPi scales other than the provisional ER scale (cf. Gough,
1962; Jackson, 1967). These procedures resulted in the final set of 29
items: Hems 331, 363, 366, and 448 (true scored) and Items 12,15,16, 3
With respect to sex differences, previous published research does
31, 38, 46, 59, 74, 111, 147, 177, 188, 205, 238, 243, 248. 258, 328, not point toward any consistent differences in mean levels of ER. In the
365, 374, 385. 416, 418, 428, and 454 (false scored; item numbers in present study, men and women did not differ in ER, either on the CAQ-
Gouglvs, 1987, 472-item version of the CPI). 4 Examples of ER items based criterion score or on the CPI-based self-report scale. Moreover,
are provided in Appendix B. The mean of the final ER scale was 21 for there were no sex differences in the alpha reliabilities of the scale, in the
both women and men. In the present studies, the ER scale was scored factor structure of ER, or in the relations between ER and the ACL
from existing CPI protocols. However, the scale is short, and the 29 scales examined in Sludy 2.
items can easily be administered as an individual scale (e.g., as part of a 4
A reviewer pointed out that the CPI shares many items with the
larger assessment battery).
Minnesota Multiphasic Personality Inventory 2 (MMPI-2; Hathaway
& McKinley, 1989) and that it might therefore be possible to score the
ER scale from the MMPI-2 as well. Item comparisons showed that 14
Results and Discussion of the 29 items were almost identical; their MMPI-2 item numbers were
Internal Consistency and Convergent and Incremental 9, 52 (reversed), 56 (reversed), 71 (reversed), 73 (reversed), 95, 130
(reversed), 160, 243 (reversed), 275 (reversed), 325 (reversed), 364
Validity ofthe Self-Report Scale (reversed), 454 (reversed), and 463 (reversed). An additional 8 MMP1-
2 items were judged by the author and one colleague to be sufficiently
To examine how well the scale performed, I analyzed its psy-
similar in meaning to be included in an MMPI-2 version of the scale;
chometric characteristics in the two independent cross-valida- these items were 38 (reversed), 167 (reversed), 328 (reversed), 350,
tion samples. One indicator of successful scale construction is 372, 391 (reversed), 516 (reversed), and 554 (reversed). In the IPSR
provided by a measure of internal consistency. The self-report sample, participants completed a combined CPI-MMPI item pool that
scale showed excellent internal consistency, with alpha coeffi- contained 19 of the proposed 22 MMPI-2 items for the ER scale. Pre-
cients of .88 and .81 in the Mills and Radcliffe samples, liminary analyses of this somewhat shorter version of the MMPI-2 scale
respectively. showed that it was internally consistent (a = .79) and very similar to
The crucial information, however, concerns validity: the degree the 29-item CPI-based ER scale (r - .84).
 CONSTRUCT OF EGO-RESILrENCY 1071

even after the joint effects of all other significant CPI scales had contrast to an anxious, neurotic, self-handicapping preoccupa-
been controlled; the partial correlations were .25, .41, and .49 tion with the negative. Thus, the first component of ER can be
in the IPSR, Mills, and Radcliffe samples, respectively. said to involve themes of confident optimism, as exemplified by
In sum, the ER self-report scale showed good internal consis- CAQ items such as "Has social poise and presence" (vs. "Tends
tency and substantial convergent validity. Furthermore, the to ruminate and have preoccupying thoughts"). Representative
scale consistently was the strongest predictor of the observer cri- CPI items loading on this first component were "My daily life is
terion and showed incremental validity above and beyond the full of things that keep me interested" and "The future seems
standard CPI scales. hopeless to me" (false).
Component 2, labeled productive and autonomous activity, in-
Internal Structure ofER via Exploratory Factor volves productivity, persistence in the face of adversity, initiative,
Analysis and independence. The CAQ descriptor defining this component
was "Is productive, gets things done" (as opposed to "Gives up and
What are the underlying components of ER? To explore the withdraws from frustration, adversity"). Examples of CPI items
components that underlie ER, I first examined its internal from the corresponding component were "Sometimes I just can't
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

structure via an exploratory factor analysis of the 26 most de- seem to get going" (false) and "I have a tendency to give up easily
fining items of the CAQ ER prototype. Initial analyses were first when I meet difficult problems" (false).
run separately by sample and by sex. Because the resulting pat- Component 3 reflects the capacity for close relationships and
terns of findings did not differ in any substantive way, the results for being insightful and socially perceptive. This component was
reported here combined all three samples; thus, the most stable labeled interpersonal warmth and insight. Defining CAQ items of
and reliable factor solution was obtained. I expected that a con- this component were "Has warmth, is compassionate" and "Has
struct as broad as ER would have multiple distinct facets, and insight into own motives and behavior." Items from the corre-
the goal of the exploratory factor analysis was to provide infor- sponding CPI component were "I must admit I have a bad temper
mation about these specific components of ER. once I get angry" (false) and "A strong person doesn't show his [ or
However, even though ER is conceptualized as combining a herl emotions and feelings" (false).
number of meaningfully distinguishable facets, the Blocks view it Finally, the fourth CAQ component was labeled skilled expres-
as a unitary construct, and its facets should therefore be correlated siveness. This component reflects an expressive interpersonal ori-
with each other. Accordingly, I performed a principal-components entation, being at ease in social settings, and being skilled in in-
analysis with an oblimin rotation on the 26 CAQ items defining teracting with others. The CAQ item defining this component was
ER. On the basis of inspection of the eigenvalues and the scree "Is skilled in social techniques" (vs. "Is emotionally bland"). Cor-
plot, four components were extracted, accounting for 61% of the responding CPI items were "It is very hard for me to tell anyone
total variance. The loadings on the four obliquely rotated compo- about myself" (false) and "When in a group of people I have trou-
nents of the 26 CAQ items are presented in Table 1. ble thinking of the right things to talk about" (false).
An examination of the rotated components revealed that four In sum, this exploration of the internal structure of the Blocks'
distinct facets, or content domains, can be identified within the CAQ definition of ER revealed four major components of ER:
general domain of ER. To interpret the conceptual meaning of confident optimism, productive and autonomous activity, inter-
these components along with the conceptual themes captured by personal warmth and insight, and skilled expressiveness. Further-
the items of the ER self-report scale, I performed a separate prin- more, analytical and conceptual correspondences were obtained
cipal-components analysis with oblimin rotation on the items of between the four pairs of components extracted from the observer
the ER scale. Once again, four components were extracted on the and the CPI-based self-report measures. This replication across
bases of an examination of the eigenvalues and the scree plot.5 To different data sources and different instruments lends credibility
test whether the respective four-component solutions obtained for to these factors as central components of ER.
the two data sources showed convergence, I computed a quantita-
tive index of correspondence. In particular, I intercorrelated the Testing the Homogeneity and Internal Structure ofER
scores retained for each of the components extracted for the CAQ
Through Confirmatory Factor Analysis
with those retained for the CPI.
The resulting correlation matrix is shown in Table 2. These To examine whether the Blocks' operationalization ofER via
correlation coefficients reflect correspondences among brief observer CAQs indeed forms a unitary construct, I used con-
facets derived from different data sources (observer vs. self- firmatory factor analysis to test the viability of a number of
reports) and different instruments with nonoverlapping item competing structural models of the construct. If ER is indeed a
content; thus, the absolute size of the correlations can be ex- broad but nevertheless unitary construct, a model that specifies
pected to be small. Note the particular pattern of convergent distinct yet correlated latent factors should provide a superior
and discriminant correlations: The values on the diagonal are fit over competing models that do not. To test this notion, I used
consistently higher than those not on the diagonal, indicating LISREL 7 (Joreskog & Sorbom, 1989) to examine the viability
that each of the CAQ components converges uniquely with just of three competing models of the internal structure of the 26
one CPI component. Given the observed pattern of one-to-one CAQ items defining ER. 6 The first model specified the existence
correspondences between the CAQ and the CPI components,
their respective conceptual meanings are discussed together. 5
The component loadings for the ER self-report scale are available
As shown in Table 1, the first component describes an opti- from the author.
6
mistic, positive, and energetic outlook and approach to life, in T would like to thank Bill Meredith for his statistical advice.
 1072 KXOHNEN

Table 1
Exploratory Factor Analysis of the 26 Items Defined as Most Characteristic and Most
Uncharacteristic of the California Adult Q-Set (CAQ) Ego-Resiliency Prototype
Oblique factor
Confident Productive Insight and Skilled
Q-set item Optimism Activity Warmth Expressiveness
Has social poise and presence .69 .01 .11 .17
Responds to humor .61 -.39 -.04 .14
Calm, relaxed in manner .59 -.02 .14 -.54
Is basically anxious (R) -.81 -.13 -.05 .09
Is vulnerable, fearful (R) -.76 -.28 .03 -.04
Tends to ruminate (R) -.76 -.10 .14 -.02
Feels cheated, victimized by life (R) -.64 -.21 -.19 .01
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Maladaptive under stress (R) -.58 -.34 -.26 .05

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Is self-defeating (R) -.56 -.42 -.21 -.02

Feels a lack of personal meaning (R) -.50 -.33 -.19 -.01
Is productive; gets things done .21 .65 .08 -.28
Values own independence .10 .64 -.18 .08
Gives up; withdraws from adversity (R) -.22 -.64 -.03 -.28
Has warmth; is compassionate .06 -.28 .80 -.01
Insight into own motives and behavior -.21 .31 .74 .12
Arouses liking and acceptance .42 -.33 .60 -.14
Perceptive of interpersonal cues .17 -.09 .59 .23
Sees to heart of problems .05 .46 .55 -.02
Is dependable and responsible .01 .26 .47 -,56
Denies unpleasant experiences (R) .07 -.14 -.54 -.40
Over-reactive to frustrations (R) -.42 .16 -.44 .17
Is skilled in social techniques .45 -.22 .10 .57
Initiates humor .53 -.14 -.08 .50
Is emotionally bland (R) -.14 -.01 -.08 -.79
Does not vary roles (R) .11 -.17 -.19 -.70
Uncomfortable with uncertainty (R) -.13 -.39 -.21 -.52
Note. Some CAQ items were abbreviated or paraphrased from J. Block (1991). N = 558. Thirteen of the
26 items were rated as extremely or quite characteristic; the remaining 13 items were rated as extremely or
quite uncharacteristic (cf. J. Block. 1991). Rotated item loadings above .40 are shown in boldface. R =
reversed.

of four uncorrelated latent factors representing the four facets plies that ER is not a unitary construct but, rather, is multidi-
obtained in the exploratory factor analysis. More specifically, mensional, with distinct components that are unrelated to each
all of the primary loadings and those cross loadings greater than other. The second model specified four correlated latent factors.
or equal to .40 that were obtained via the exploratory factor This model reflects the hypothesis that all four latent factors
analysis were freed to be estimated by LISREL. This model im- are related to each other because they all tap aspects of one
overarching construct. Finally, the third model tested this hier-
archical notion explicitly: It specified one general factor un-
derlying the four correlated latent factors, which in turn were
Table2 expected to account for the relations among the ER items. This
Intercorrelations Among Factor Scores Retainedfor the Four model was essentially a "second-order" factor structure that
Factor Solutions Obtained in Separate Principal-Components specified one higher order factor (i.e., a general ER factor) un-
Analyses of the California Adult Q-Set (CAQ) and California derlying the relations among the four lower order factors. Both
Psychological Inventory (CPl) Ego-Resiliency Measures the correlated factors model and the second-order factor model
CAQ factor are consistent with the present conceptualization of ER as uni-
tary and should thus provide approximately equally goodfitsto
CPI factor 1 2 3 4 the data. The uncorrelated factors model, in contrast, is hypoth-
esized to provide a significantly worse fit than either of the other
] .28 .05 -.04 -.12 two models.
2 .01 .23 .10 -.01
3 .01 -.05 .23 .11 The LISREL results indicate that the uncorrelated factors
4 .05 .08 -.03 .29 model showed clearly the worst fit across all three standard fit
Note. N = 541. Correlations above .20 are shown in boldface. Corre- indexes (i.e., the chi-square statistic, goodness-of-fit index, and
lations exceeding .11 were significant at least atp < .01. Factor labels root mean square of the residuals). The two models that speci-
indicate the order in which they appeared in the oblique rotation. fied variants of the "general factor" notion showed better and
 CONSTRUCT OF EGO-RESILIENCY 1073

acceptable levels of fit. Given that all three models were nested represent the four factorial components of ER. Before inter-
within each other (i.e., the parameters of one model were a sub- preting these results, it should be noted that the absolute
set of the other), it is possible to compare the models statisti- magnitude of the relations of ER with the ACL scales is less
cally. The difference in Iheir chi-square values was distributed important than the observed pattern of convergent and dis-
as a chi-square value with (df{ - df2) degrees of freedom criminant relations. The emphasis on the pattern of relations
(Bentler & Bonett, 1980). When this technique was applied to is necessary because extraneous influences (such as shared
evaluate the relative goodness offitof the three models, both of method variance, attenuation due to unreliability, and sam-
the general factor models provided a statistically better fit than ple fluctuations) are likely to affect the magnitude of the cor-
the uncorrelated factors model, x 2 (6, N= 558) = 1,386.0,/? < relations. Note, for example, that the correlations involving
.001,and x 2 (3,JV=558)= 1,388.7,/><.001 (for the compar- the self-report scale tended to be higher than those obtained
isons with the correlated and higher order factor models, for the observer measure, a finding due to the shared method
respectively). Moreover, the correlated factors model and the variance between the self-reports. To obtain a more reliable
second-order factor model did not show significantly different estimate of the strength of the observed relations, I averaged
fits, x 2 (3, N = 558) = 2.72, ns. In sum, thefindingsfrom the the observed correlation coefficients across the middle four
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

confirmatory factor analysis add to those from the exploratory columns in Table 3 (i.e., across samples and across ER
This document is copyrighted by the American Psychological Association or one of its allied publishers.

factor analysis by providing support for the notion that ER is measures) .8 The last column in the table provides the average
best conceived of as a superordinate yet unitary personality re- correlation of ER with each of the ACL scales.
source that combines a number of important and more specific An examination of the four vectors of correlations in Table 3
facets of personality. shows that Ihe relative magnitude of correlations for the ob-
server measure and the self-report scale across the two samples
Study 2: Testing the Pattern of Convergent and was quite similar. To quantify the degree of similarity between
Discriminant Relations of ER With Specific the patterns of correlations, I computed the average intercorre-
Personality Attributes lation among the first four vectors of correlation coefficients
shown in Table 3 (i.e., the mean intercorrelation of the four
On the basis of the Blocks* theory and thefindingsconcerning middle columns). The mean intercorrelation of .91 was im-
the internal structure of ER obtained in Study 1,1 expected ER pressively high and suggests considerable generalizability of the
to be related to certain personality attributes but not to others. findings both across the two ER measures and across the differ-
To examine the predicted pattern of convergent and discrimi- ent samples.
nant relations, I related ER to clusters of scales drawn from a
The convergent relations of ER with the four ACL clusters
widely used personality inventory, the ACL (Gough & Heil-
can be summarized as follows. High scorers on the scales chosen
brun. 1983). I expected convergent relations with clusters that
to represent confident optimism tend to be enterprising and
reflected the four factorial components obtained in Study 1 and
adaptive individuals with a positive and optimistic outlook and
discriminant relations with ACL scales measuring aspects of
with solid feelings of personal worth. They also have the capa-
ego-control and impulse expression.
bilities to initiate activities and to carry them through. The pro-
ductive activity cluster of scales is indicative of effectiveness in
Method setting goals and of an ambitious, persistent, and directed ap-
Participants proach to attaining them. In addition, high-scoring individuals
possess a sense of self-sufficiency and are competent and re-
The analyses in this study used the IPSR (N = 350) and Mills (N =
102) samples previously described in Study 1 (total A' = 452).
sourceful. Despite their independence, however, these individu-
als also have the ability to accept and follow rules of social living,
are self-disciplined, and are willing to fulfill their duties and ob-
Measures
ligations. High scorers on the interpersonal warmth and insight
Measures of ER. The CAQ observer measure and the CPl-based scale cluster have the capacity for close relationships, affiliative
self-report scale of ER were used in this study. Information about both warmth, and supportive relating to others. They also have the
measures is provided in Study 1. desire and the capacity to understand themselves and others and
ACL. The ACL (Gough & Heilbrun. 1983) is a widely used 300-
show resourceful and efficient use of their intellectual abilities.
item personality inventory that provides scores on 37 specific scales. To
test ER's convergent validity, I selected scales from the ACL and
Also implied by these scales are tolerance of ambiguity and
grouped them into four content domains that corresponded to the four openness to experience. High scorers on the skilled expressive-
components found to underlie ER. ? To test evidence for discriminant ness cluster are assertive, self-assured, and comfortable in social
validity, I chose 5 additional scales to capture aspects of ego-control and settings. They also display interpersonal versatility and sponta-
impulse expression. neity and are adroit in interacting with and directing the behav-
iors of others.
Results and Discussion
Does ER Show the Predicted Convergent and 7
It is important to note that there is no exact one-to-one correspon-
Discriminant Relations With Personality Attributes dence between the trait domains and the four components of ER ob-
Measured by Clusters of ACL Scales? tained in Study 1. The scales for each cluster were chosen to best ap-
proximate the content domain of the four factorial components.
As can be seen in Table 3, ER showed substantial con- 8
Fisher r-\o-z transformations of the correlations were performed
vergent relations with the clusters of ACL scales selected to before they were summed and averaged.
 1074 KLOHNEN

Table 3
Correlations of Observer-Based and Self-Reported Ego-Resiliency With Adjective Check List
(ACL) Scales Selected to Represent the Four Factorial Components
of Ego-Resiliency in Two Samples
IPSR sample Mills sample

ACL scale Observer Self-report Observer Self-report Average

Confident optimism
Self-confidence .40 .65 .35 .61 .51
Personal Adjustment .27 .53 .36 .62 .45
Ideal Self .25 .57 .32 .54 .43
Productive activity
Achievement .19 .49 .20 .36 .31
Leadership .13 .39 .21 .50 .31
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Succorance -.29 -.51 -.25 -.55 -.41

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Interpersonal insight and warmth

Nurturance .11 .26 .29 .48 .29
Intraception .11 .35 .27 .54 .33
Creative Personality .37 .51 .32 .42 .42
Skilled expressiveness
Affiliation .21 .51 .33 .64 .44
Dominance .31 .60 .29 .46 .42
Abasement -.25 .54 -.28 -.47 -.39

Note. JYS for the analyses were 102 and 350 for the Mills and IPSR samples, respectively. Average corre-
lations were based on summing and averaging the r-Xo-z transformed correlations for each ACL scale across
the four vectors. All correlations exceeding | .251 are shown in boldface. Across samples, correlations ex-
ceeding .20 were significant at least at p < .05. IPSR - Institute of Personality and Social Research.

However, as hypothesized, ER was not related to all possible adult adjustment index for men and that it is based on data that
domains of personality. In line with J. H. Block and Block's reflect the life outcomes of adult participants in their mid-40s.
(1980) theorizing, the discriminant relations with thefiveACL Thus, this measure provides an index of effective functioning
scales selected to tap aspects of ego-control and impulse expres- and adjustment that can be "observed" in the participants' lives
sion were all negligible in size. Across the two measures of ER in a variety of domains of experience. The adjustment index
and the two samples, the mean intercorrelations were - . 17 for includes the work, relationship, and family domains, as well as
Self-Control,. 10 for Order, .01 for Aggression,. 13 for Change, physical and psychological health. In addition, I tested whether
and - . 13 for Deference.9 the ER scale shows incremental validity in predicting adjust-
In summary, this set of correlations confirmed the hypothe- ment beyond standard CP1 scales. To test for the incremental
sized pattern of convergent and discriminant relations. The validity of the ER scale, I performed stepwisc multiple regres-
convergent relations hetween ER and the ACL scales chosen to sions as well as partial correlation analyses.
tap its four components provide an exlernal validation of the
internally derived components of ER. The discriminant re- Method
lations provide support for the Blocks' theorizing about the in-
dependence of ER from various aspects of ego-control. Finally, Participants
it is important to note that the findings held for both the self- The analyses in this study used the Mills (A' = 75) and Radcliffe (<V
report scale and the observer measure. Thus, the observed re- = l03)samples described in Study I (total A'= 178).
sults cannot simply be due to response sets on the part of the
participants. Measures
Study 3: Test of the External Validity of ER: Predicting ER. The CAQ observer measure and the CPl-based self-report scale
of L:R were used in this study. Information about both measures is pro-
Adjustment in Daily Life vided in Study 1.
Study 3 tested one of the most central theoretical proposi- Index of Adult Adjustment This adjustment index (Picano, 1989)
tions about ER: Does ER have implications for adjustment and is a multifaccted instrument that measures levels of effective function-
ing and adult adjustment among women. It closely follows Vaillant's
effective functioning in everyday life? To provide evidence for
(1977) Adult Adjustment Scale developed for men. The index is based
the criterion validity of ER, I examined its relations with a com- on life data and reliably coded questionnaire material. It is composed of
prehensive index of adult adjustment (Picano, 1989) that was
available in two samples of women. It is important to note that
this index was developed as an equivalent of Vaillanf s (1977) ' A complete table is available from the author.
 CONSTRUCT OF EGO-RESIUENCY 1075

20 indicators that represent empirical markers of effective functioning one analysis did one of the three remaining predictors explain a
likely to be observed in the lives of adult women. In addition to an over- significant amount of additional variance of the criterion.
all adjustment score, the measure provides subscores on work and social
Does ER show incremental validity beyond standard CPI
adjustment and on physical and psychological health. Examples of spe-
cific domains included in the index are work attainment, years of stable
scales? To examine this issue, I tested whether ER significantly
marriage, health problems, and alcohol and drug use (see Picano, 1989, predicts adjustment after accounting for the influence of the
for a more detailed description of the index). Scores on the adjustment three competing predictors. Across both samples, ER signifi-
index were available for the Mills and Radcliffe samples. It should be cantly predicted both global and domain-specific adjustment af-
noted that the adjustment index was used somewhat differently in the ter the effects of all three competing predictors had been par-
two samples because the available data differed somewhat in the tialed out.
samples. The significant and replicated relations of ER with effective
functioning and adjustment in a variety of life domains provide
Results and Discussion strong evidence for the hypothesized link between ER and ad-
justment. Moreover, thefindingsheld for both the observer and
Does ER Predict Adjustment and Effective Functioning? the self-report instruments, thus showing evidence of the con-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

struct validity of both measures and ensuring that the obtained

This document is copyrighted by the American Psychological Association or one of its allied publishers.

To test the hypothesis that ER predicts effective functioning relations cannot simply be due to response biases or shared
and adjustment in a number of life domains, I correlated scores method variance. Finally, the incremental validity of the ER
on the observer and self-report measures of ER with the Index scale over standard CPI scales in predicting midlife adjustment
of Adult Adjustment. Table 4 shows these correlations in two illustrates the utility and uniqueness of the new scale.
samples. As predicted, both observer-based and self-reported
ER were significantly related to overall adjustment. Moreover,
the effect held in both samples. General Discussion
In addition, Table 4 shows that the two measures of ER were The present set of studies explored the nature of the Blocks'
significantly related to both subdomams of adjustment: the construct of ER and showed that measures of this personality
work and social domain and the psychological and physical resource predict adjustment and effective functioning in adult-
health domain. Again, these effects held in both samples. The hood. Exploratory and confirmatory factor analyses suggest
finding that the correlations with the two subdomains of adjust- that ER is a broad and superordinate construct that combines
ment were significant and similar in size suggests that the con- a number of distinct personality attributes. In particular, ego-
ception of adjustment predicted by ER is not specific to only resilient individuals have a sense of active and meaningful en-
one area of life; instead, it includes both intrapsychic and inter- gagement with the world. Their positive and energetic approach
personal components, as one would predict from the analysis of to life is grounded in confident, autonomous, and competent
the factorial components of ER in Study 1. functioning and a sense of mastery within a wide range of life
domains. In addition, ego-resilient individuals are perceptive
Does the ER Scale Show Incremental Validity Over and insightful and have the capacity for warm and open re-
lations with others. They also possess the necessary interper-
Standard CPI Scales in Predicting Adjustment? sonal skills and social poise to effectively negotiate the social
To test the incremental validity of the ER scale, I performed world.
stepwise multiple regressions using as "competing predictors" I also found ER to be related to measures of four domains of
the three CPI scales that correlated most highly with the ER adaptive personality functioning suggested by the factor analy-
scale in all samples. Those three scales were Well-Being, Self- ses: psychological well-being, productivity, effective intellectual
Realization, and Independence. In both samples and for all functioning and perceptiveness, and an adroit and open inter-
three adjustment criteria, ER was indeed the first and thus personal style. Furthermore, in line with the Blocks' conceptu-
strongest predictor to enter the equation. Furthermore, in only alization of the construct, I found ER to be unrelated to ego-

Table 4
Test ofthe Link Between Ego-Resiliency and Adjustment: Correlations of the Observer and
Self-Report Measures ofEgo-Resiliency in Two Samples of Women at Midlife
Observer-based Self-reported
ego-resiliency ego-resiliency

Adjustment index Mills Radcliffe Mills Radcliffe

Global adjustment .35* .53** .49* .23*

Work and social .32* .50** .34* .23*
Psychological and physical health .21* .41** ,39* .20*

Note. N=75 for the Mills sample and N = 103 and I • for the observer and self-report measures of ego-
resiliency, respectively, in the Radcliffe sample.
* p < .05. **p < .01 (one-tailed test).
 1076 KLOHNEN

control and impulse expression. Most important, the ER mea- the scale consisted of equal numbers of women and men, and
sures had substantial external validity; They predicted adjust- the analyses of this sample did not identify any substantive
ment and effective functioning in adulthood measured in terms differences between the sexes (see Footnote 3).
of adaptive outcomes across various life domains. Overall, these Another potential limitation concerns the imbalanced keying
findings provide evidence for a conceptualization of ER as a of the items on the ER scale. Most of the items are keyed in
personality resource that enables individuals toflexiblymodify the direction of ego-brittleness rather than ER. This is likely to
their impulse expression and thus adaptively respond to envi- reflect the greater salience and observability of behaviors indic-
ronmental contingencies and shape them in accordance with ative of maladjustment or brittleness, as compared with behav-
their personal goals and desires (cf.J. H. Block & Block, 1980). iors indicative of emotional stability (e.g., John & Robins,
1993, p. 539). If so, this imbalance in keying may well have a
A New Self-Report Measure ofER substantive basis. In fact, given the substantial validity coeffi-
cients obtained in the present studies, the imbalanced keying
In addition to the aim of enhancing understanding of the na- apparently does not pose a threat to the validity and utility of
ture and definition of ER, a second major aim was to introduce the scale. Furthermore, thisfindingis consistent with the long-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

a new self-report measure of the construct. The findings indi- standing recognition that acquiescence is not an important re-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

cate that the scale has much promise as a measure of ER when sponse bias (J. Block, 1965;Rorer, 1990).
the labor-intensive and resource-intensive CAQ-based observer
measure is not available or cannot be obtained; The scale has
ER and Other Conceptualizations ofAdjustment and
good internal consistency reliability, substantial convergent va-
lidity with J. H. Block and Block's (1980) observer-based mea- Well-Being
sure of ER, and impressive external criterion validity with Pi- How does J. H. Block and Block's (1980) ER constructfitin
cano's (1989) index of adjustment. Furthermore, the scale the generalfieldof psychological health, well-being, and adjust-
showed incremental validity above and beyond existing CPI ment? Over the years, numerous definitions of psychological
scales in predicting both the observer-based ER criterion and health have been proposed (e.g., Allport, 1961; Jahoda, 1958;
the adjustment index. Moreover, all of thefindingswere repli- Korchin, 1976). Although these definitions differ somewhat in
cated effects obtained in at least two samples of adults. These terms of their breadth and emphasis, they seem to agree on sev-
results demonstrate the utility and uniqueness of the ER scale eral fundamental attributes characteristic of psychologically
in addition to the standard CPI scales. well-adjusted individuals: (a) the ability to be happy and
The ER scale can be scored from the CPI, which is one of the contented with a sense of direction and purpose; (b) the capac-
most widely used personality inventories. As such, the CPI has ity for productive work and a sense of competence and environ-
been included in many existing archival data sets and longitudi- mental mastery; (c) emotional security, self-acceptance, self-
nal studies (e.g., J. Block, 1971; Cartwright & Wink, 1994; Hel- knowledge, and a realistic and undistorted perception of one-
son et al,, 1984; Vandewater & Stewart, in press), making these self, others, and one's surroundings; and (d) interpersonal
valuable data sources amenable to research on ER. The 29-item adequacy and the capacity for warm and caring relating to oth-
self-report scale can also be administered separately as an indi- ers and for intimacy and respect. These general criteria for men-
vidual scale, making it possible to study the effects of ER in a tal health and effective functioning, I suggest, resemble the com-
wide variety of new research contexts. ER has already been ponents of ER uncovered in the present research: (a) confident
shown to play an important role in a large number of research optimism, (b) autonomous and productive activity, (c) inter-
domains (e.g., Asendorpf & van Aken, 1991; Colvin, 1993; personal insight and warmth, and (d) skilled expressiveness. It
Funder & Block, 1989; Kobak & Sceery, 1988; Sroufe et al., thus appears that the construct and operation alization of ER
1993). Using the self-report scale, researchers can now study tap and encompass many of the aspects thought essential to psy-
these and other effects of ER both in archival data sets and in chological health and adjustment.
new or ongoing research projects. So far, I have argued that there are important correspon-
Some promising work is already under way. For example, us- dences between ER and common definitions of psychological
ing archival data, Klohnen, Vandewater, and Young (in press) health. Can any evidence be found that the four components of
demonstrated that ER predicts the way in which women expe- ER are important in empirical research as well? One relevant
rience and renegotiate the midlife transition, a potentially line of research that has been conducted independently of work
stressful period of reflection, reevaluation, and change. In com- on ER has investigated the attributes that are consistently found
parison with the rest of the sample, ego-resilient women were to be characteristic of stress-resistant or resilient children (e.g.,
more satisfied, were more likely to further their education or Garmezy, 1985; Werner & Smith, 1982). Among these attri-
continue to build a career, and showed fewer health-related butes are reflectiveness and confidence when confronting new
problems. situations, activity level, cognitive skills, and positive respon-
One of the limitations of the present research is that the two siveness to others. These characteristics of children, 1 propose,
samples used to examine the prediction of adjustment in mid- resemble the four components found to underlie ER in adults
life included only women. Future research needs to examine in the present research. Apparently, the major components that
in more detail the validity of the ER scale in samples of men. define ER in adulthood are similar to those that independent
However, there is little reason to expect pervasive sex differences research has shown to characterize stress-resistant children.
in the meaning, the measurement, or the validity of the con- Going beyond research on attributes of resilient children, nu-
struct. Note that the large derivation sample used to construct merous important and useful constructs have been articulated
 CONSTRUCT OF EGO-RESILIENCY 1077

in the vast literature on psychological health (e.g., Marshall, Obviously, further research is necessary before there can be
Wortman, Vickers, Kusulas, & Hervig, 1994). How might ER a full account of the nature of ER and its relations to other
be related to these other conceptions of psychological well-being conceptions of psychological health and adjustment. In general,
and adjustment (e.g., Ryff, 1989)? The present research has I believe that careful research and theorizing is needed to better
emphasized two important aspects of ER that need to be con- understand the numerous conceptualizations and measures of
sidered when approaching this question: First, ER is denned psychological health and adjustment and their interrelations
and measured at a broad level of abstraction, and, second, it and overlap. Eventually, such research will elucidate the funda-
includes four conceptually distinct components. As I suggested mental aspects of human adaptability. However, these funda-
earlier, ER provides a comprehensive account of psychological mentals of human adaptability are likely to be obscured by the
health and adjustment and, thus, should be related to other, plethora of diversely labeled measures, each designed to capture
more narrowly denned conceptions of psychological well-being. somewhat different but nonetheless overlapping definitions of
For example, dispositional optimism (Scheier & Carver, 1985) well-being and adjustment. The present studies explored the na-
is defined by the general expectation that desirable outcomes ture of ER as operationalized via the CAQ and a self-report scale
are likely to occur in the future. Given the component structure and explicated hypotheses about the relations between ER and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

of ER, one would expect that dispositional optimism is related other relevant constructs. Future research needs to test these
This document is copyrighted by the American Psychological Association or one of its allied publishers.

to ER and that this relation is due primarily to one of the four hypotheses empirically. The focus in the present research was
components of ER, namely that of confident optimism. There on ER because it is a superordinate construct that is theoreti-
are other constructs in the psychological health literature that cally derived and conceptually well grounded. As such, ER
should be related to ER because they focus on some but not all seems well suited to contribute to an understanding of individ-
of the components of ER. For example, constructs that empha- ual differences in human adaptability.
size the content domains captured by the components of con-
fident optimism and productive and autonomous activity in-
clude locus of control (Rotter, 1966), self-mastery (Pearlin &
References
Schooler, 1978), and self-efficacy (Bandura, 1982). Allport, G. W. (1961). Pattern and growth in personality. New \brk:
Holt, Rinehart & Winston.
Another interesting personality concept likely to be related to Asendorpf, J. B., & van Aken, M. A. (1991). Correlates of the temporal
ER is social intelligence (e.g., Cantor & Kihlstrom, 1987; Kos- consistency of personality patterns in childhood. Journal of Person-
mitzki & John, 1993). Kosmitzki and John (1993) provided a ality, 59, 689-703.
synthesis of thefindingson implicit and explicit conceptions of Bandura, A. (1982). Self-efficacy mechanism in human agency. Amer-
social intelligence and found that social intelligence is best con- ican Psychologist, 37, 122-147.
ceived as a combination of cognitive components (e.g., under- Bentler, P. M., & Bonett, D. G. ( 1980). Significance tests and goodness
standing people, perspective taking, and knowing social rules) and of fit in the analysis of covariance structures. Psychological Bulletin,
behavioral components (e.g., being good at dealing with people, 88, 588-606.
social adaptability, and interpersonal warmth). Consequently, I Block, J. {1950). An experimental investigation of the construct ofego-
suggest that social intelligence should be related to ER and overlap controi Unpublished doctoral dissertation, Stanford University, Stan-
ford, CA.
most closely with two of its components: interpersonal warmth
Block, J. (1965). The challenge oj response sets. New York: Applelon-
and insight and skilled expressiveness. Century-Crofts.
Two other important constructs also predicted to relate to ER Block, J. (1971). Lives through lime. Berkeley, CA: Bancroft.
are positive and negative emotionality. Positive emotionality en- Block, J.( 1978). The Q-sort method in personality assessment and psy-
compasses behavioral and temperamental characteristics con- chiatric research. Palo Alto, CA: Consulting Psychologists Press,
ducive to joy, excitement, and vigor and to states of positive (original work published 1961)
engagement, whereas negative emotionality is associated with Block, J. (1991). Prototypes for the California Adult Q-Set. Berkeley:
anxiety, anger, and related states of negative engagement (e.g., Department of Psychology, University of California.
Tellegen, 1985; Watson & Clark, 1984). Both of these dimen- Block, J.H.( 1951). An experimental study oftopological representation
sions are expected to be related to ER, with positive emotional- of ego structure. Unpublished doctoral dissertation, Stanford Univer-
sity, Stanford, CA.
ity related positively and negative emotionality related
Block, J. H., & Block, J. (1980). The role of ego-control and ego-resil-
negatively. iency in the organization of behavior. In W. A. Collins (Ed.), Minne-
Some preliminary evidence for these predictions comes from sota symposia on child psychology (Vol. 13, pp. 39-101). Hillsdale,
Klohnen's (1994) finding that ER is related primarily to three NJ: Erlbaum.
of the Big Five personality dimensions (John, 1990; McCrae, Cantor, N., & Kihlstrom, J. F. (1987). Personality and social intelli-
Costa, & Busch, 1986): high extraversion, low neuroticism, and gence. Englewood Cliffs, NJ: Prentice Hall.
high conscientiousness. Note that extraversion has been inter- Cartwright, L. K., & Wink, P. (1994). Personality change in women
preted as positive emotionality and that neuroticism has been physicians from medical student years to mid-40s. Psychology of
interpreted as negative emotionality (e.g., Costa & McCrae, Women Quarterly 18, 291-308.
1980; Tellegen et al., 1988; Watson, Clark, & Tellegen, 1988). Caspi, A.; & Herbener, E. S. (1990). Continuity and change: Assortative
marriage and the consistency of personality in adulthood. Journal of
Thus, within both the Big Three (Tellegen, 1985) and the Big
Personality and Social Psychology 58, 250-258.
Five (John, 1990) frameworks, ER reflects both high positive Colvin, C. R. (1993). Childhood antecedents of young-adult judgabil-
emotionality and low negative emotionality, along with appro- ity. Journal of Personality, 61, 611-635.
priate conscientiousness to perform effectively in achievement Compas, B. E. (1987). Coping with stress during childhood and adoles-
and work-related settings. cence. Psychological Bulletin, 101, 393-403.
 1078 KLOHNEN

Costa, P. T., & McCrae, R. R. (1980). Influence of extraversion and the middle years: Ego-resiliency and successful midlife adjustment in
neuroticism on subjective well-being: Happy and unhappy people. women. Psychology and Aging.
Journal of Personality and Social Psychology, 38. 668-678. Kobak, R. R., & Sceery, A. (1988). Attachment in late adolescence:
Cowen, E. L., Wyman, P. A., Work, W. C , & Parker, G. R. (1990). The Working models, affect regulation, and representations of self and
Rochester Child Resilience Project: Overview and summary of first others. Child Development, 59, 135-146.
year findings. Development and Psychopathology, 2, 193-212. Korchin, S. (1976). Modern clinical psychology New York: Basic
Egeland, B., Carlson. E., & Sroufe, L. A. (1993). Resilience as process. Books.
Development and Psychopathology, 5, 517-528. Kosmitzki, C , & John, O. P. (1993). The implicit use of explicit con-
Funder, D. C , & Block, J. (1989). The role of ego-control, ego-resil- ceptions of social intelligence. Personality and Individual Differences,
iency, and IQ in delay of gratification in adolescence. Journal of Per- 15, 11-23.
sonality and Social Psychology, 57, 1041-1050. Marshall, G. N., Wortman, C. B., Vickers, R. R., Kusulas, J. W., &
Funder, D. C , &Colvin, C. R. (1988). Friends and strangers: Acquain- Hervig, L. K. (1994). The five-factor model of personality as a frame-
tanceship, agreement, and the accuracy of personality judgment. work for personality-health research. Journal of Personality and So-
Journal of Personality and Social Psychology 55, 149-158. cial Psychology 67, 278-286.
Garmezy, N. {1985). Stress resistant children: The search for protective Masten, A. S., Best, K. M., & Garmezy, N. (1990). Resilience and de-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

velopment: Contributions from the study of children who overcome

This document is copyrighted by the American Psychological Association or one of its allied publishers.

factors. In J. Stevenson (Ed.), Recent research in developmentalpsy-

chopathology (Suppl. No. 4, pp. 213-233). Oxford, England: Perga- adversity. Development and Psychopathology, 2, 425-444.
mon Press. McCrae, R. R., Costa, P. T., & Busch, C. M. (1986). Evaluating com-
Garmezy, N. (1991). Resiliency and vulnerability to adverse develop- prehensiveness in personality systems: The California Q-Set and the
mental outcomes associated with poverty. American Behavioral Sci- five-factor model. Journal of Personality, 54, 430-446.
entist, 34, 416-430. Ozer, D. J., & Gjerdc, P. F. (1989). Patterns of personality consistency
Gough, H. G. (1957). Manual for the California Psychological Inven- and change from childhood through adolescence. Journal of Person-
tory. Palo Alto, CA: Consulting Psychologists Press. ality, 57, 483-507.
Gough, H. G. (1962). Imagination—Undeveloped resource. In Pearlin, L. I., & Schooler, C. (1978). The structure of coping. Journal
5. J. Parnes&H. F. Harding (Eds.), A source book for creative think- of Health and Social Behavior, 19, 2-21.
ing (pp. 217-226). New York: Scribner's. (Original work published Peterson, B. E., & Klohnen, C. E. (1995). The realization ofgenerativ-
1957) ity in two samples of women at midlife. Psychology and Aging, 10,
Gough, H. G. (1987). Administrator's guide for the California Psycho- 20-29.
logical Inventory. Palo Alto, CA: Consulting Psychologists Press. Picano, J. (1989). Development and validation of a life history index of
Gough, H. G., & Heilbrun, A. B., Jr. (1983). The Adjective Check List adult adjustment for women. Journal of Personality Assessment, 53,
manual Palo Alto, CA: Consulting Psychologists Press. 308-318.
Hathaway, S. R., & McKinley, J. C. (1989). MMPI-2: Manual for ad- Rarer, L. G. (1990). Personality assessment: A conceptual survey. In
ministration and scoring. Minneapolis: University of Minnesota L. A. Pervin (Ed.), Handbook of personality: Theory and research
(pp. 693-720). New York: Guilford Press.
Press.
Helson, R. (1967). Personality characteristics and developmental his- Rotter, J. B. (1966). Generalized expectancies for internal versus exter-
nal control of reinforcement. Psychological Monographs, 80{ 1,
tory of creative college women. Genetic Psychology Monographs, 76,
Whole No. 609).
205-256.
Rutter, M. (1979). Protective factors in children's responses to stress
Helson, R., Mitchell, V., & Moane, G. (1984). Personality and patterns
and disadvantage. In M. W Kent & J. E. Rolf (Eds.), Primary pre-
of adherence and non-adherence to the social clock. Journal of Per-
vention of psychopathology: Social competence in children (Vol. 3, pp.
sonality and Social Psychology, 46, 1079-1096.
49-74). Hanover, NH: University Press of New England.
Helson, R., & Wink, P. (1992). Personality change in women from the
Rutter, M. (1987). Psychosocial resilience and protective mechanisms.
early 40s to early 50s. Psychology and Aging, 7, 45-55.
American Journal of Orthopsychiatrics, 57, 316-331.
Jackson, D. N. (1967). Personality Research Form manual. Goshen,
Ryff, C. D. (1989). Happiness is everything, or is it? Explorations on
NY: Research Psychologists Press.
the meaning of psychological well-being. Journal of Personality and
Jahoda, M. (1958). Current concepts of positive mental health. New- Social Psychology. 57, 1069-1081.
York: Basic Books. Scheier, M. F., & Carver, C. S. (1985). Optimism, coping, and health:
John, O. P. (1990). The "Big Five" factor taxonomy: Dimensions of Assessment and implications of generalized outcome expectancies.
personality in the natural language and in questionnaires. In Health Psychology, 4, 219-247.
L. A. Pervin (Ed.), Handbook of personality: Theory and research Sroufe, L. A., Carlson, E., & Shulman, S. (1993). Individuals in rela-
(pp. 66-100). New York: Guilford Press. tionships: Development from infancy through adolescence. In
John, O. P., & Robins, R. W. (1993). Determinants of interjudge D. C. Funder, R. D. Parke, C. Tomlinson-Keasey, & K. Widaman
agreement on personality traits: The Big Five domains, observability, (Eds.), Studying lives through lime: Approaches to personality and
evaluativeness, and the unique perspective of the self. Journal of Per- development (pp. 315-342). Washington, DC: American Psychologi-
sonality, 61, 521-551. cal Association.
Jdreskog, K. G., & Sorbom, D. (1989). LISREL 7: A guide to the pro- Stewart, A. J. (1978). A longitudinal study of coping styles in self-de-
gram and applications (2nd ed.). Chicago: SPSS. fining and socially defined women. Journal of Consulting and Clini-
Klohnen, E. C. (1994, August). Definition and developmental anteced- cal Psychology, 46, 1079-1084.
ents of Block's concept of ego-resiliency. In O. P. John (Chair), Stewart, A. J. (1980). Personality and situation in the prediction of
Trails, types, and stages—Integration of personality models through women's life patterns. Psychology of Women Quarterly, 5, 195-206.
longitudinal research. Symposium conducted at the 102nd Annual Stewart, A. J., & Salt, P. (1981). Life stress, life-styles, depression and
Convention of the American Psychological Association, Los Angeles, illness in adult women. Journal of Personality and Social Psychology,
CA. 40, 1063-1069.
Klohnen, E. C . Vandewater, E. A., & Young, A. (in press). Negotiating Stewart, A. J., & Vandewater E. (1993). The Radcliffe class of 1964:
 CONSTRUCT OF EGO-RESILIENCY 1079
Career and family social clock projects in a transitional cohort. In career commitments and personality development. In M. E. Lach-
K. D. Hulbert & D. Tickton Schuster (Eds.), Women's lives through man & S. B. James (Eds.), Multiple paths of midlife development.
time: Educated American women of the twentieth century (pp. 235- Chicago: University of Chicago Press.
258^ San Francisco: Jossey-Bass. Watson, D,, & Clark, L. A. (1984). Negative affectivity: The disposition
Tellegen, A. (1985). Structures of mood and personality and their rele- to experience negative emotional states. Psychological Bulletin, 96,
vance to assessing anxiety, with an emphasis on self-report. In 465-490.
A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and val-
{pp. 681 -716). HilJsdale, NJ: Erlbaum. idation of brief measures of positive and negative affect: The PANAS
Tellegen, A., Lykken, D. T, Bouchard, T. J., Wilcox, K. J., Segal, N. L., scales. Journal of Personality and Social Psychology, 54, 1063-1070.
& Rich, S. (1988). Personality similarity in twins reared apart and Werner, E. E;, & Smith, R. S. (1982). Vulnerable but invincible: A study
together. Journal of Personality and Social Psychology, 54, 1031 — of resilient children. New York: McGraw-Hill.
1039. York, K., & John, O. (1992). The four faces of Eve: A typological anal-
Vaillant, G. E. (1977). Adaptation to life. Boston: Little, Brown. ysis of women's personality at midlife. Journal of Personality and So-
Vandewater, E. A., & Stewart, A. S. {in press). Changes in women's cial Psychology, 63.494-508.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Appendix A
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Thirteen Most Characteristic and Thirteen Most Uncharacteristic Items of the California Adult Q-Set Ego-Resiliency
Criterion Definition by J. Block (1991)
The top 13 items were rated as extremely or guile characteristic, the bottom 13 items were rated as
extremely or quite uncharacteristic (cf. J. Block, 1991).

Most Characteristic Most Uncharacteristic

60. Has insight into own motives and behavior 45. Brittle ego-defense; maladaptive under stress
35. Has warmth; capacity for close relationships 55. Is self-defeating
92. Has social poise and presence 9. Is uncomfortable with uncertainty and complexities
26. Is productive; gets things done 34. Over-reactive to minor frustrations, irritable
33. Calm, relaxed in manner 86. Denies unpleasant thoughts and experiences
15. Is skilled in social techniques of imaginary play 100. Does not vary roles; relates to all in same way
64. Socially perceptive of interpersonal cues 68. Is basically anxious
83. Able to see to the heart of important problems 30. Gives up and withdraws from frustration, adversity
2. Is genuinely dependable and responsible person 97. Is emotionally bland
56. Responds to humor 40. Is vulnerable to real or fancied threat; fearful
96. Values own independence and autonomy 79. Tends to ruminate and have pre-occupying thoughts
28. Tends to arouse liking and acceptance 78. Feels cheated and victimized by life
18. Initiates humor 22, Feels a lack of personal meaning in life

Appendix B

Examples of Representative Items From the Self-Report Ego-Resiliency Scale Based

on the California Psychological Inventory
The 29 items on the ego-resiliency scale were drawn from the California Psychological Inventory (Gough,
1987).

I often feel as if the world was just passing me by. (false) I feel like giving up quickly when things go wrong, (false)
Most of the time I feel happy. People seem naturally to turn to me when decisions have to be made.
I am certainly lacking in self-confidence, (false) It is hard for me to tell anyone about myself, (false)
My daily life is full of things that keep me interested.

Received November 18, 1994

Revision received August 3, 1995
Accepted August 14, 1995