Child Development, May/June 2000, Volume 71, Number 3, Pages 573–575

Research on Resilience: Response to Commentaries

Suniya S. Luthar, Dante Cicchetti, and Bronwyn Becker

Clarifications are provided with respect to two sets of issues raised in preceding commentaries. First, interac-
tion effects are undoubtedly salient in resilience research; yet main effect findings can be equally critical from
an intervention perspective. Second, although resilience research and prevention science reflect similar broad
objectives, the former (but not usually the latter) involves explicit attention to positive adjustment outcomes in
addition to the avoidance of psychopathology.

INTRODUCTION cance as one that can be harnessed to help children

whose parents are clinically depressed.
In a series of commentaries encompassing diverse
To borrow from von Eye and Schuster (2000), in in-
themes, Robinson (2000), Roosa (2000), and von Eye and
vestigating the effects of a particular “protective factor”
Schuster (2000) have provided constructive directions
in a two group design—exposed/not exposed to
for future work on resilience and have also raised
risk—there are four possible patterns of findings, in
some additional questions regarding the distinguish-
terms of whether effects for that factor are present
ing features of research in this area. We appreciate the
(), as opposed to absent or weak (). These pat-
opportunity to clarify those distinguishing features.
terns are as follows:
In the following response, we consider two major is-
sues that have been raised: the degree to which resil- 1. Risk exposed  Not exposed 
ience research is defined by particular statistical 2. Risk exposed  Not exposed 
models (i.e., those involving interaction effects), and 3. Risk exposed  Not exposed 
the overlap with, versus distinctness from, preven- 4. Risk exposed  Not exposed 
tion research.
Considering these patterns of results, resilience in-
vestigators are interested not in interaction effects per
INTERACTION EFFECTS AS “DEFINING se (pattern 2 is of little relevance) but in all findings of
FEATURES” OF THE FIELD OF RESILIENCE substantial effects within the risk group under con-
sideration. Without question, of central concern are
As Roosa (2000) suggests, interaction effects are cer-
interactions establishing that a construct is potent in
tainly at the heart of resilience research. Detection of
the risk-exposed group but relatively trivial in the
such interaction effects is what sensitized scholars in
not-exposed group (pattern 1). On the other hand,
this field to the fact that protective and vulnerability
equally pertinent are those interactions that have
forces are often highly context-specific, with modest or
strong beneficial effects across diverse groups (pat-
negligible effects in the absence of adversity but pow-
tern 3)—instances that von Eye and Schuster (2000, p.
erful ones in the presence of particular risk conditions.
564) characterize as having an odds ratio of 1, where
The substantial relevance of interaction effects to
the “moderator is equally effective across the two
resilience research does not, however, imply that
groups.”
main effects are irrelevant; from an intervention per-
Our comments on the statistical instability of inter-
spective, information conveyed by main effect find-
action effects are not intended to devalue their impor-
ings is equally important. For the scientist who seeks
tance in resilience research (see Roosa, 2000); rather,
to foster positive adaptation among a group that typ-
they are to caution against undue emphasis on those
ically shows negative outcomes (e.g., children of de-
effects. Given the limits of currently available statisti-
pressed mothers), the central task is to determine
cal techniques and of research designs (e.g., as a result
what allows some of these youngsters to do well. The
of “inappropriate” comparison groups, as Roosa sug-
fact that the identified “protective factor” might also
gests), it would be a mistake for resilience researchers
help other groups in no way diminishes its signifi-
to conclude that they had identified nothing pertinent
Reply to commentaries on Luther, Cicchetti, & Becker, “The
Construct of Resilience: A Critical Evaluation and Guidelines for © 2000 by the Society for Research in Child Development, Inc.
Future Work,” by von Eye & Schuster, Roosa, and Robinson. All rights reserved. 0009-3920/2000/7103-0005
574 Child Development

to protective factors if interaction terms in their anal- promotion as an objective, citing the reason that,
yses were statistically nonsignificant but main effects “. . . health promotion is not driven by an emphasis on
explained a high proportion of variance. illness, but rather by a focus on the enhancement of
Assuming that interaction effects are critical yet do well-being, i.e., to enhance competence, self-esteem,
not exclusively define resilience research, what does and a sense of well-being, rather than to intervene to
render this area distinctive? From our perspective, prevent psychological and social problems or mental
the strongest single distinguishing feature lies in the disorders” (Mrazek & Haggerty, 1994, p. 27).
focus on positive indicators among groups typically It is critical to note that resilience researchers’ em-
thought of in terms of their problems. In their earliest phasis on wellness by no means implies their dis-
writings, founders of the field of resilience, such as missal of that which does not approach excellence.
Norman Garmezy and Emmy Werner, forced devel- The focus, simply, is on trajectories that are “unex-
opmentalists to amend a long-standing focus on defi- pectedly positive,” that is, those where outcomes are
cits among at-risk groups. They argued for explicit at- substantially better than one might expect on the ba-
tention to the strengths of risk-exposed individuals as sis of how most individuals are affected by the adver-
well, both in terms of adjustment outcomes (compe- sity in question. As we have noted in our review pa-
tence in addition to symptomatology) and in terms of per (Luthar, Cicchetti, & Becker, 2000), often it is most
characteristics that promote positive adaptation— meaningful to operationalize positive outcomes in
assets or protective factors, as well as “liabilities” or terms of the avoidance of serious symptomatology,
vulnerability-enhancing ones. such as antisocial behaviors or internalizing prob-
lems. In any situation where the risk condition is con-
ceptually and statistically linked with psychiatric di-
RESILIENCE RESEARCH AND
agnoses (i.e., high odds ratios), it would be entirely
PREVENTION RESEARCH
logical to treat evasion of psychopathology, rather
Undoubtedly, there are several substantive similari- than the maintenance of high social competence, as
ties between resilience research and prevention re- the appropriate indicator of resilient adaptation.
search (Robinson, 2000), but there are also some differ- On the subject of stringency of standards applied
ences. Researchers in both fields seek to understand in conferring labels of resilience, a note of clarification
what helps some people “do well” and ultimately to is in order with regard to illustrations that Robinson
apply this information in interventions. A major dis- (2000) has provided from the Luthar (1991) study. In
tinction, however, lies in the connotations of the the latter investigation, to earn the classification of
phrase “doing well.” For preventionists, this usually “positive adaptation,” students were required to have
represents disease avoidance (Coie et al., 1993; Mrazek scores in the top 16% of the distribution within that
& Haggerty, 1994; Munoz, Mrazek, & Haggerty, 1996; particular inner-city sample and not in the top 10% of
Robinson, 2000). Resilience researchers, on the other any national norms. Thus, they were judged in terms
hand, are concerned with wellness in addition to the of the school-based behaviors of other children with
absence of dysfunction. Scholars in this tradition are largely similar ecological life settings rather than
committed to the notion of maximizing potential and against any absolute standards.
well-being among at-risk individuals and not just to the To be sure, the cutoff of 1 standard deviation that
evasion of serious psychiatric disorder (Cowen, 1999). Luthar (1991) used is an arbitrary one; other cutoffs
This difference in overall foci is vividly captured in such as the top tertile or the top quartile can be, and
a chapter by Cowen (1999), which includes citations have been used. Decisions about where the bar should
from diverse documents on primary prevention. For be set (Robinson, 2000) must be dictated by the degree
example, a report by a Task Panel of nine leaders in to which the researcher is interested primarily in the
prevention science (Coie et al., 1993, p. 1013) indi- evasion of problems among at-risk youth as opposed to
cated that the overarching goals of this discipline are their achievement of clear-cut competence. While these
“. . . to prevent or moderate major human dysfunc- themes are entirely complementary, any efforts to pro-
tions” and to “eliminate or mitigate the causes of dis- mote them within intervention programs would re-
order.” An annotated bibliography of over 1,300 pri- quire attention to distinct target domains, as well, fre-
mary prevention articles (Trickett, Dahiyat, & Selby, quently, the use of different strategies (Cowen, 1999).
1994) indicated that in 98% of the items included, the
focus was on disease prevention and only 2% dealt
SUMMARY
with mental health promotion (Dinges, 1994). Fur-
thermore, the 1994 Institute of Medicine Report on In response to commentaries on our review of resil-
preventive intervention explicitly excluded health ience research, we offer clarifications regarding two
 Luthar, Cicchetti, and Becker 575

sets of questions about distinguishing features of REFERENCES

work in this area. With regard to statistical findings
Coie, J. D., Watt, N. F., West, S. G., Hawkins, J. D., Asarnow,
that “define” resilience research, interaction effects J. R., Markman, H. J., Ramey, S. L., Shure, M. B., & Long,
are undoubtedly of great importance: they sensitize B. (1993). The science of prevention: A conceptual frame-
researchers to the fact that many so-called protective work and some directions for a national research pro-
or vulnerability processes can be highly potent in the gram. American Psychologist, 48, 1013–1022.
context of particular risk conditions, even though Cowen, E. L. (1999). In sickness and in health: Primary pre-
their effects are trivial in the absence of adversity. vention’s vows revisited. In D. Cicchetti & S. L. Toth
The importance of such effects does not, however, (Eds.), Rochester Symposium on Developmental Psychopa-
detract from the significance of main-effect associa- thology: Vol 9. Developmental approaches to prevention and
tions. Any evidence that a construct is strongly intervention (pp. 1–24). Rochester, NY: University of
Rochester Press.
linked with adaptation in a given at-risk group is of
Dinges, N. (1994). Mental health promotion. In P. J. Mrazek
high relevance for those who seek to intervene with
& R. J. Haggerty (Eds.), Reducing risks for mental disorders:
that group. Frontiers for preventive intervention (pp. 333–355). Wash-
With regard to resilience vis-à-vis prevention re- ington, DC: National Academy Press.
search, a salient distinguishing feature lies in the Luthar, S. S. (1991). Vulnerability and resilience: A study of
emphasis on positive outcomes among resilience high-risk adolescents. Child Development, 62, 600–616.
researchers, which contrasts with the attention to Luthar, S. S., Cicchetti, D., & Becker, B. E. (2000). The con-
disease prevention typically identified with the struct of resilience: A critical evaluation and guidelines
field of prevention science. It is important to note for future work. Child Development, 71, 543–562.
that emphasis on positive adaptation in resilience Mrazek, P. J., & Haggerty, R. J. (1994). Reducing risks for men-
does not imply rigid requirements of excellence in tal disorders: Frontiers for preventive intervention research.
Washington, DC: National Academy Press.
any absolute sense but rather denotes a focus on
Munoz, R. F., Mrazek, P. J., & Haggerty, R. J. (1996). Institute
trajectories that reflect unusually positive adapta- of Medicine report on prevention of mental disorder:
tion, given what usually occurs within the adversity Summary and commentary. American Psychologist, 51,
under consideration. 1116–1122.
Robinson, J. L. (2000). Are there implications for prevention
research from studies on resilience? Child Development,
ADDRESSES AND AFFILIATIONS 71, 570–572.
Roosa, M. W. (2000). Some thoughts about resilience versus
Corresponding author: Suniya S. Luthar, Department of
positive development, main effects versus 269, and the
Human Development, Teachers College, Box 133, Co-
value of resilience. Child Development, 71, 567–569.
lumbia University, 525 West 120th Street, New York, NY Trickett, E. J., Dahiyat, C., & Selby, P. (1994). Pimary preven-
10027-6696; e-mail: Suniya.Luthar@columbia.edu. tion in mental health: An annotated bibliography. Rockville,
Dante Cicchetti is at Mt. Hope Family Center, University MD: U.S. Department of Health and Human Services.
of Rochester, Rochester, NY; and Bronwyn Becker is at von Eye, A., & Schuster, C. (2000). The odds of resilience.
Teachers College, Columbia University, New York, NY. Child Development, 71, 563–566.