Journal of Personality Disorders, 19(5), 466–486, 2005
2005 The Guilford Press
THE CHILDREN IN THE COMMUNITY
STUDY OF DEVELOPMENTAL COURSE
OF PERSONALITY DISORDER
Patricia Cohen, PhD, Thomas N. Crawford, PhD,
Jeffrey G. Johnson, PhD, and Stephanie Kasen, PhD
The Children in the Community (CIC) Study is an ongoing investigation
of the course of psychiatric disorders including personality disorders
(PDs) in an epidemiological sample of about 800 youths. In addition to
tracking developmental trajectories over 20 years from adolescence into
adulthood, the CIC Study has used prospective data to investigate early
risks for Axis II disorders and symptoms (including both environmental
factors and early characteristics), implications of comorbidity with Axis
I disorders, and associated negative prognostic risk of adolescent PDs
into adulthood. The substantial independent impact of PD on subse-
quent Axis I disorders, suicide attempts, violent and criminal behavior,
interpersonal conflict, and other problematic adult outcomes confirms
the importance of attention to these problems when they manifest in
early adolescence. The implications of study findings for potential
changes in the DSM are discussed.
The CIC Study began as a follow-up to a 1975 study of a large random
sample of children (ages 1–10 years) living in households in 100 residen-
tial areas sampled in 2 upstate New York counties. The original study was
designed to assess the level of need for children’s services and validate
social indicators of that need (Kogan, Smith, & Jenkins, 1977). Data came
from maternal interviews and covered a wide range of developmental, tem-
perament, health, and environmental variables. When first followed-up in
1983, study goals shifted to focus on predictors of Axis I psychiatric disor-
ders in early adolescence (mean age = 14) (see Cohen & Cohen, 1996 for a
full description of sampling method, retention, and characteristics). Axis I
disorders were assessed in interviews with children and mothers using the
Diagnostic Interview Schedule for Children (DISC-1; Costello, Edelbrock,
Dulcan, Kalas, & Klaric, 1984). The protocol also covered theoretically and
empirically plausible risks for disorders assessed with multiple measures
of family, peer, neighborhood, and school environment. Health, personal-
From Columbia University and New York State Psychiatric Institute.
Address correspondence to Patricia Cohen, Ph.D., 100 Haven Ave. Apt. 31F, New York, NY
10032; E-mail: firstname.lastname@example.org.
CHILDREN IN THE COMMUNITY STUDY467
ity, attitudes, values, and behaviors were reported by both parent and
From this first assessment Cohen was interested in exploring the devel-
opment of PDs in children and adolescents. However, despite DSM-III rec-
ognition of likely early life origins of PDs, there were no existing age-appro-
priate PD measures for adolescents.
To avoid lengthening the interview with unnecessary repetition, we used
plausible symptom indicators already available in the protocol. Most of
these child- and parent-reported items covered the “normal” range as well
as extremes on traits or behaviors that approximated some Axis II symp-
toms (e.g., distrusting others, difficulty controlling anger, reckless disre-
gard for safety, abandonment fears, etc.). To provide coverage of remaining
symptoms, Cohen selected items from the Personality Disorder Question-
naire (PDQ; Hyler, Rieder, Spitzer, & Williams, 1982) with age-appropriate
revisions when needed. A few new items were written when no alternative
seemed age-appropriate. DISC-1 items were not used in PD scales in order
to ensure independent assessment of Axis I and II disorders.
The CIC sample was reassessed in mid-adolescence (mean age = 16).
Preparation for this follow-up permitted Drs. Cohen and Schwab-Stone to
augment the previous protocol with additional items better suited to the
PD criteria, especially as revised for the DSM-III-R. New items were
adapted from a prototype of the Structured Clinical Interview for DSM-III-
R Personality Disorders (Spitzer & Williams, 1986). Bernstein reviewed all
non-DISC youth- and parent-reported items to create PD symptom scales
and diagnostic algorithms. This work led to the first CIC publications on
PD (Bernstein, Cohen, Velez, Schwab-Stone, Siever, & Shinsato, 1993;
Bezirganian, Cohen, & Brook, 1993) based on the best information avail-
able in early adolescence and more complete data from the mid-adolescent
The next data collection took place in early adulthood (mean age = 22)
and was largely modeled on the earlier protocols. Although data on the
developmental course of PDs was potentially available from age 9 (the
youngest participants in the first assessment) to age 27 (the oldest in the
young adult assessment), investigation of age changes in symptom levels
required consistent measurement in repeated assessments. Furthermore,
the DSM-IV came out with several changes in diagnostic criteria. These
considerations led us to revise our item selection and algorithms. Working
with Cohen, Jeffrey G. Johnson reviewed the parent and youth items in
the protocol and produced consistent symptom scales and algorithms in
all three assessments that covered almost all DSM-IV criteria. Diagnoses
were made using scaled diagnostic criteria counts and standard DSM-IV
The reassessment of the CIC sample at mean age = 33 did not include
468 COHEN ET AL.
maternal interviews, thus preventing use of the existing PD scales. New
measures were developed by Thomas Crawford based on self-report items
alone. All items were part of early and later adult assessments (mean ages
22 and 33) and thus constituted a consistent measure of PD across the 11
years. At age 33 a clinical assessment of PDs using the Structured Clinical
Interview for DSM-IV Personality Disorders (SCID-II; First, Gibbon, Spitzer,
Williams, & Benjamin, 1997) permitted evaluation of concordance between
the CIC measures and the more standard assessment. Clusters A and B
diagnoses based on CIC algorithms both demonstrated higher concor-
dance with SCID-II diagnoses than most comparable self-report instru-
ments (Crawford et al., 2005). Concordance for Cluster C diagnoses was
closer to the published average. Unlike other self-report instruments, algo-
rithms for the CIC Study do not overestimate diagnoses when compared
with SCID-II clinical diagnoses. Scaled measures of PD showed surpris-
ingly good prediction of CIC and SCID-II scales and associated disability
at mean age 33 from CIC measures recorded 11 years earlier.
As would be expected, there is substantial correlation between these dif-
ferent versions of our measures. The most significant changes came when
adolescent and young adult measures combining youth and parent re-
ports were compared to measures using self-report alone. Agreement be-
tween corresponding PD scales was still high (mean r = .63) and higher at
the cluster level (mean r = .72) even though parent and youth reports are
often discordant. We thus have found little reason to be concerned about
problems in generalizing findings from one PD version to another.
However, two related problems have become apparent over time that are
not unique to this study or our assessment instruments. The first pertains
to arbitrary criterion-based cutoffs demanded by the DSM. Despite the
utility of diagnoses for clinical and policy purposes, we agree that on the
whole diagnostic cutoffs have little empirical justification. Future taxo-
nomic and latent class work by our group or others may examine potential
empirical justification for such thresholds. A related problem occurs when
individuals do not meet criteria for any single PD but nevertheless have
many symptoms from a combination of PDs. Our analyses show that the
adverse outcomes for these people with personality disorder not otherwise
specified (PDNOS) are as serious as those in people who met full diagnostic
criteria for one or more PDs (Johnson et al., in press).
The second problem is determining just when a given criterion has been
met. Because symptom items have graded response options (e.g., always,
most of the time, sometimes, never), there are multiple choices about
which item cut-points will best define when a diagnostic criterion has been
met. Response extremity and relative low endorsement rates are logically
expected for a behavior or other manifestation to be considered a symptom
in the general population. However, definitions of “extremity” or “low rate
endorsement” will vary from one diagnostic algorithm to another. These
decisions can have substantive effects on the conclusions drawn from an
empirical investigation. For instance, extremity of cutoff can have large
CHILDREN IN THE COMMUNITY STUDY 469
effects on the correlations with the other variables in an analysis. We have
found that this problem is minimized by analyses using scaled symptom
measures as well as categorical diagnoses, with attention to consistency
across empirically based conclusions.
Another issue pertains to the DSM-IV’s requirement that adolescent PD
symptoms persist for at least 1 year. Most PD items in our assessment and
other self-reports do not specify duration, but are phrased as descriptions
of the person and presumed reasonably stable. Nevertheless, some of our
analyses of long-term outcomes consider adolescents with PDs to be only
those meeting or nearly meeting criteria at both mean ages 14 and 16.
SUBSTANTIVE QUESTIONS AND METHODOLOGICAL ISSUES
Given prospective data accumulated over 20 years, the CIC Study is
unique in being able to address questions about developmental trajecto-
ries of adolescent PDs. The Study’s comprehensive assessment of Axis I
and II disorders has also permitted investigation of the developmental sig-
nificance of comorbidity between these two sets of disorders. Comorbidity
estimates based on our community sample are especially valuable because
they are not inflated by ascertainment bias typically associated with re-
ferred samples (Berkson, 1946). One of the drawbacks, however, is that
community samples yield relatively few cases of individual PDs, thereby
limiting statistical power. To increase power we often analyze PDs grouped
into diagnostic clusters, generally checking to ensure that findings are not
Prospective data have permitted investigation of adult outcome variables
(e.g., Axis I and II disorders, social impairment, suicide, violence, etc.) that
can be predicted by adolescent PDs. Moreover, CIC data have been used
to show how early risk factors (e.g., childhood abuse, neglect, maladaptive
parenting, school experiences) predict PDs in adolescence and adulthood.
Insofar as prospectively reported data have been used, our findings were
not distorted by recall bias that undermines confidence in retrospective
reports. Analyses also generally include consideration of probable or es-
tablished correlated risks as covariates. When testing effects for a given
PD, we have learned the importance of controlling for other co-occuring
disorders. For example, adolescent PD initially appeared to be predictive
of early adult eating and weight problems (Johnson, Cohen, Kotler, Kasen,
& Brook, 2002). However, its effect disappeared once adolescent depres-
sive disorder and earlier eating problems were added to the model.
STUDY FINDINGS RELATED TO AGGREGATED PD SYMPTOMS
Change and Stability in Symptom Levels
We have shown that mean PD symptoms overall and virtually all disorders
are consistently highest in early adolescence and are followed by an essen-
tially linear decline from age 9 to 27 (Johnson, Cohen, Kasen et al., 2000).
484 COHEN ET AL.
Berkson, J. (1946). Limitations of the appli-
cation of fourfold table analysis to hos-
pital data. Biometrics, 2, 47–53.
Bernstein, D. P., Cohen, P., Skodol, A., Bezir-
ganian, S., & Brook, J. S. (1996).
Childhood antecedents of adolescent
personality disorders. American Jour-
nal of Psychiatry, 153, 907–913.
Bernstein, D. P., Cohen, P., Velez, C. N.,
Schwab-Stone, M., Siever, L. J., &
Shinsato, L. (1993). Prevalence and
stability of the DSM III-R personality
disorders in a community-based sur-
vey of adolescents. American Journal
of Psychiatry, 150, 1237–1243.
Bezirganian, S., Cohen, P., & Brook, J.
(1993). The impact of mother-child in-
teraction on the development of bor-
derline personality disorder. American
Journal of Psychiatry. 150, 1836–1842.
Brooks-Gunn, J., & Kirsch, B. (1984). Life
events and boundaries of midlife for
women. In G. Baruch & J. Brooks-
Gunn (Eds.), Women in midlife (pp.
11–30). New York: Plenum Press.
Caron, C., & Rutter, M. (1991). Comorbidity
in child psychopathology: Concepts,
issues and research strategies. Jour-
nal of Child Psychology and Psychia-
try, 32, 1063–1080.
Chen, H., Cohen, P., Johnson, J. G., Kasen,
S., Sneed, J., R., & Crawford, T. N.
(2004). Adolescent personality disor-
der and conflict with romantic part-
ners during the transition to adult-
hood. Journal of Personality Disorders,
Cohen, P. (1996). Childhood risks for young
adult symptoms of personality disor-
der: Method and substance. Multivari-
ate Behavioral Research, 31, 121–148.
Cohen, P., Brown, J. & Smailes, E. (2001).
Child abuse and neglect and the devel-
opment of mental disorders in the gen-
eral population. Developmental Psy-
Cohen, P., Chen, H., Crawford, T. N., Gor-
don, K., & Brook, J. S. (2004). Person-
ality disorders in early adolescence
and the development of later substance
use disorders in
ulation. Manuscript under review.
Cohen, P., Chen, H., Kasen, S., Johnson,
J. G., Crawford, T. N., & Gordon K.
(2005). Adolescent Cluster A personal-
ity disorder symptoms, role assump-
tion in the transition to adulthood,
and resolution or persistence of symp-
toms. Development and Psychopathol-
ogy, 17(2), 549–568.
Cohen, P., & Cohen, J. (1996). Life values
and adolescent mental health. Mah-
wah, NJ: Lawrence Erlbaum Associates.
Cohen, P., & Crawford, T. N. (2005). Develop-
mental issues. In J. M. Oldham, A. E.
Skodol, & D. Bender (Eds.), Textbook
of Personality Disorders (pp. 171–185).
Washington, DC: American Psychiat-
Costello, A. J., Edelbrock, C. S., Dulcan,
M. K., Kalas, R., & Klaric, S. H. (1984).
Testing of the NIMH Diagnostic Inter-
view Schedule for Children (DISC) in a
clinical population: Final report to the
Center for Epidemiological Studies, Na-
tional Institute for Mental Health. Pitts-
burgh: University of Pittsburgh.
Cowan, C. P., & Cowan, P. A. (1992). The
landmark ten-year study: When part-
ners become parents. The big life change
for couples. New York: Basic Books.
Crawford, T. N., Cohen, P., & Brook, J. S.
(2001a). Dramatic-erratic personality
disorder symptoms: I. Continuity from
early adolescence into adulthood. Jour-
nal of Personality Disorders, 15, 319–
Crawford, T. N., Cohen, P., & Brook, J. S.
(2001b). Dramatic-erratic personality
disorder symptoms: II. Developmental
pathways from early adolescence to
adulthood. Journal of Personality Dis-
orders, 15, 336–350.
Crawford, T. N., Cohen, P., Johnson, J. G.,
Kasen, S., First, M. B., Gordon, K., &
Brook, J. S. (2005). Self-reported per-
sonality disorder in the Children in the
Community sample: Convergent and
prospective validity in late adolescence
and adulthood. Journal of Personality
Disorders, 19(1), 30–52.
Crawford, T. N., Cohen, P., Johnson, J. G.,
Sneed, J. R., & Brook, J. S. (2004). The
course and psychosocial correlates of
personality disorder symptoms in ado-
lescence: Erickson’s developmental the-
ory revisited. Journal of Youth and Ad-
olescence, 33, 373–387.
Erikson, E. H. (1950). Childhood and society.
New York: Norton.
the general pop-
CHILDREN IN THE COMMUNITY STUDY 485
Erikson, E. H. (1968). Identity: Youth and cri-
sis. New York: Norton.
Ferro, T., Klein, D. N., Schwartz, J. E.,
Kasch, K. L., & Leader, J. B. (1998).
30-month stability of personality dis-
order diagnoses in depressed outpa-
tients. American Journal of Psychiatry,
First, M. B., Gibbon, M., Spitzer, R. L., Wil-
liams, J. B. W., & Benjamin, L. S.
(1997). User’s guide for the Structured
Clinical Interview for the DSM-IV Per-
sonality Disorders. Washington, DC:
American Psychiatric Press.
Hamigami, F., McArdle, J. J., & Cohen,
P.(2000). A new approach to modeling
bivariate dynamic relationships ap-
plied to evaluation of DSM-III person-
ality disorder symptoms. In V. J. Molf-
ese & D. L. Molfese (Eds.), Temperament
and personality development across
the life span (pp. 253–280). Mahwah,
NJ: Lawrence Erlbaum Associates.
Hyler, S. E., Rieder, R., Spitzer, R., & Wil-
liams, J. (1982). The Personality Diag-
nostic Questionnaire (PDQ). New York:
New York State Psychiatric Institute.
Johnson, J. G., Cohen, P., Brown, J., Smailes,
E., & Bernstein, D. (1999). Childhood
maltreatment increases risk for per-
sonality disorders during young adult-
hood: Findings of a community-based
longitudinal study. Archives of General
Psychiatry, 56, 600–606.
Johnson, J. G., Cohen, P., Kasen, S., Skodol,
A. E., Hamagami, F., & Brook, J. S.
(2000). Age-related change in person-
ality disorder trait levels between early
adolescence and adulthood: A commu-
nity-based longitudinal investigation.
Acta Psychiatrica Scandinavica, 102,
Johnson, J. G., Cohen, P., Kasen, S., Smailes,
E., & Brook, J. S. (2001). The associa-
tion of maladaptive parental behavior
with psychiatric disorder among par-
ents and their offspring. Archives of
General Psychiatry, 58, 453–460.
Johnson, J. G., Cohen, P., Kotler, L., Kasen,
S., & Brook, J. S. (2002). Psychiatric
disorders associated with risk for the
development of eating disorders dur-
ing adolescence and early adulthood.
Journal of Consulting and Clinical Psy-
chology, 70, 1119–1128.
Johnson, J. G., Cohen, P., Skodol, A., Old-
ham, J. M., Kasen, S., & Brook, J.
(1999). Personality disorders in adoles-
cence and risk of major mental disor-
ders and suicidality during adulthood.
Archives of General Psychiatry, 56, 805–
Johnson, J. G., Cohen, P., Smailes, E.,
Kasen, S. Oldham, J. M., Skodol, A. E.,
& Brook, J. S. (2000). Adolescent per-
sonality disorders associated with vio-
lence and criminal behavior during ad-
olescence and early adulthood. American
Journal of Psychiatry, 157, 1406–1412.
Johnson, J. G., Cohen, P., Smailes, E. M.,
Skodol, A. E., Brown, J., & Oldham,
J. M. (2001). Childhood verbal abuse
and risk for personality disorders dur-
ing adolescence and early adulthood.
Comprehensive Psychiatry, 42, 16–23.
Johnson, J. G., First, M. B., Cohen, P., Sko-
dol, A. E., Kasen, S., & Brook, J. S. (in
press). Adverse outcomes associated
with personality disorder not other-
wise specified (PDNOS) in a commu-
nity sample. American Journal of Psy-
Johnson, J. G., Smailes, E. M., Cohen, P.,
Brown, J., & Bernstein, D. P. (2000).
Associations between four types of
childhood neglect and personality dis-
order symptoms during adolescence
and early adulthood: Findings of a
community-based longitudinal study.
Journal of Personality Disorders, 14,
Kasen, S., Cohen, P., & Brook, J. S. (1998).
dropout, adolescent pregnancy, and
young adult deviant behavior. Journal
of Adolescent Research, 13, 49–72.
Kasen, S., Cohen, P., Skodol, A., Johnson,
J. G., & Brook, J. S. (1999). The influ-
ence of child and adolescent psychiat-
ric disorders on young adult person-
ality disorder. American Journal of
Psychiatry, 156, 1529–1535.
Kasen, S., Cohen, P., Skodol, A. E., Johnson,
J. G., Smailes, E., & Brook, J. S.
(2001). Alternative pathways of conti-
nuity: Childhood depression and adult
personality disorder. Archives of Gen-
eral Psychiatry, 58, 231–236.
Kasen, S., Cohen, P., Skodol, A. E., Johnson,
J. G., Smailes, E., & Brook, J. S.
(2002). Substance use and the devel-
opment of antisocial personality in de-
pressed adolescents [Reply]. Archives
of General Psychiatry, 59, 665.
486COHEN ET AL.
Kendler, K. S., McGuire, M., Gruenberg,
A. M., O’Hare, A., Spellman, M., &
Walsh, D. (1993). The Roscommon Fam-
ily Study: III. Schizophrenia-related per-
sonality disorders in relatives. Archives
of General Psychiatry, 50, 781–788.
Kogan, L. S., Smith, J., & Jenkins, S. (1977).
Ecological validity of indicator data as
predictors of survey findings. Journal
of Social Science Research, 1, 117–132.
Lachman, M. E., & James, J. B. (1997). Mul-
tiple paths of midlife development. Chi-
cago: University of Chicago Press.
Lenzenweger, M. F. (1999). Stability and
change in personality disorder fea-
tures. Archives of General Psychiatry,
Loranger, A. (1990). The impact of DSM III
on diagnostic practice in a university
hospital: A comparison of DSM II and
DSM III in 10,914 patients. Archives of
General Psychiatry, 47, 672–675.
McCrae, R. R., Costa, P. T., Schneewind,
K. A., Bremond, J., Brune, E., Rivolier,
J., Strelau, J., Inglehart, M. R., Mar-
kus, H. et al. (1989). The self, person-
ality and social adaptation. In J. P.
Forgas & J. M. Innes (Eds.), Recent ad-
vances in social psychology: An inter-
national perspective (pp. 429–490).
Oxford, UK: North-Holland.
Moffitt, T. E. (1993). Adolescence-limited and
life-course-persistent antisocial behav-
ior: A developmental taxonomy. Psy-
chological Review, 100, 674–701.
Oldham, J. M., & Morris, L. B. (1990). The
personality self-portrait: Why you think,
work, love, and act the way you do.
New York: Bantam Books.
Paris, J. (2003). Personality disorders over
time: Precursors, course and outcome.
Journal of Personality Disorders, 17,
Quay, H. C. (1993). The psychobiology of un-
dersocialized conduct disorder: A the-
oretical perspective. Development and
Psychopathology, 5, 65–180.
Sperry, L. (2003). Handbook of diagnosis and
treatment of DSM-IV-TR personality
disorders (2nd ed.). New York: Brun-
Spitzer, R. L., & Williams, J. B. W. (1986).
Structured Clinical Interview for DSM-
III-R Personality Disorders (SCID-II).
New York: New York State Psychiatric
Trull, T. J. (1993). Temporal stability and va-
lidity of two personality disorder in-
ventories. Psychological Assessment, 5,