European Child & Adolescent Psychiatry

12:1–8 (2003) DOI 10.1007/s00787-003-0298-2 ORIGINAL CONTRIBUTION

Peter Muris The Strengths and Difficulties

Cor Meesters
Frank van den Berg Questionnaire (SDQ)
Further evidence for its reliability and validity in a
community sample of Dutch children and adolescents

■ Abstract This study was a first normal children and adolescents cally meaningful way with other
attempt to examine the psychomet- (N = 562) and their parents com- measures of psychopathology. It
ric properties of the Strengths and pleted the SDQ along with a num- can be concluded that the psycho-
Difficulties Questionnaire (SDQ) in ber of other psychopathology mea- metric properties of the parent-
Dutch youths. A large sample of sures. Factor analysis of the SDQ and self-report version of the SDQ
yielded five factors that were in were satisfactory in this Dutch
keeping with the hypothesised sub- community sample. Moreover, the
scales of hyperactivity-inattention, current data provide further sup-
Accepted: 30 September 2002 emotional symptoms, peer prob- port for the utility of the SDQ as an
lems, conduct problems, and index of psychopathological symp-
Peter Muris, Ph.D. () · C. Meesters · prosocial behaviour. Furthermore, toms in youths.
F. van den Berg internal consistency, test-retest sta-
Department of bility, and parent-youth agreement ■ Key words psychopathological
Medical, Clinical, of the various SDQ scales were ac- symptoms – children and
and Experimental Psychology
Maastricht University, P. O. Box 616 ceptable. Finally, the concurrent adolescents – Strengths and
6200 MD Maastricht, The Netherlands validity of the SDQ was good: that Difficulties Questionnaire
E-Mail: [email protected] is, its scores correlated in a theoreti-

Secondly, clinicians can employ such measures as part of

Introduction the clinical assessment in order to obtain an initial idea
about the type and severity of the psychiatric problem.
Prevalence studies have indicated that between 17 % and Thirdly and finally, standardised questionnaires are also
26 % of the children and adolescents meet the diagnos- helpful to those clinicians who want to quantify the ef-
tic criteria for at least one psychiatric disorder [30]. In fects of treatment [4].
particular, behaviour problems such as oppositional-de- Currently, the two most frequently used instruments
fiant disorder, conduct disorder and attention-deficit for measuring psychopathological symptoms in chil-
and hyperactivity disorder (ADHD) and emotional dren and adolescents are the Rutter [26, 27] and Achen-
problems such as anxiety disorders and depression are bach [1–3] questionnaires. The Rutter questionnaires
frequently diagnosed among youths [7, 30]. are brief rating scales to be completed by parents and
Questionnaires for measuring psychopathological teachers that have proven to be reliable and valid in-
symptoms in children and adolescents are important for dexes of child psychopathology in many contexts [8].
three reasons. Firstly, despite the fairly high prevalence However, developed more than three decades ago, the
rates of behaviour and emotional problems, it should be Rutter questionnaires are now somewhat outdated in
noted that only a small percentage of the children and that they do not adequately cover many areas of con-
adolescents actually come in contact with mental health temporary interest such as concentration, impulsivity,
services. Thus, questionnaires that can be used for de- victimisation, and prosocial behaviour. Furthermore,
ECAP 298

tecting youths who are at high risk for developing be- while it is generally assumed that the assessment of
havioural and emotional problems are highly relevant. child psychopathology should be based on information
2 European Child & Adolescent Psychiatry, Vol. 12, No. 1 (2003)
© Steinkopff Verlag 2003

obtained from multiple informants [4], there is no ver- ployed as an effective screen for child psychiatric disor-
sion of the Rutter questionnaire that can be completed ders in community samples [12, 13].
by children and adolescents themselves. The Achenbach The current study was a first attempt to examine the
questionnaires are better in this respect as they include psychometric properties of the SDQ in Dutch youths. A
versions for completion by parents viz. the Child Be- large sample of normal children and adolescents
haviour Checklist (CBCL), teachers viz. the Teacher Re- (N = 562) and their parents completed the new ques-
port Form (TRF), and youths themselves viz. the Youth tionnaire along with a number of other measures,
Self-Report (YSR) [1–3]. CBCL, TRF, and YSR intend to namely the Achenbach questionnaire [1, 3], the Child
cover a broad range of behavioural and emotional Depression Inventory (CDI; 22), the Revised Children’s
symptoms that are particularly relevant when assessing Manifest Anxiety Scale (RCMAS; 25), and the ADHD
children and adolescents in clinical settings. Yet, for Questionnaire (ADHDQ; 28). The following issues were
screening or research purposes, the Achenbach ques- investigated: 1) the factor structure of the SDQ, 2) the re-
tionnaires seem less useful as they are quite long and liability (internal consistency and test-retest stability) of
contain many items that are not relevant to the major- the questionnaire, and 3) the concurrent validity of the
ity of children. Moreover, researchers have had difficul- SDQ through its associations with other measures of
ties to replicate the empirically-derived syndrome di- psychopathology. As data were used from different in-
mensions of the Achenbach scales [18] and to relate formants, viz. youths and parents, it became also possi-
them to current conceptualisations of child psy- ble to study 4) the parent-youth agreement of the SDQ.
chopathology [23].
Thus, while the Rutter and Achenbach questionnaires
are well-respected and useful instruments for assessing Method
child psychopathology, both seem to have their own
weaknesses. With this in mind, Goodman [9] started to ■ Participants and procedure
revise the Rutter questionnaire by updating its content
and by adding items about children’s strengths, a Twelve-hundred-and-fourteen children and adolescents
process that eventually led to the development of the from seven regular primary and secondary schools and
Strengths and Difficulties Questionnaire (SDQ). The their parents were invited to participate in the present
SDQ is a brief measure covering the most important cur- study. More than half of the youths and their parents
rent domains of child psychopathology (i. e. emotional (N = 622; i. e. 51.2 %) responded favourably to our invi-
symptoms, conduct problems, hyperactivity-inatten- tation and completed the informed consent form. Youth
tion, and peer problems) as well as personal strengths participants were asked to complete the set of question-
(i. e. prosocial behaviour) that can be completed by par- naires at school during regular classes, with a teacher
ents, teachers, and youths themselves. The psychometric and a research assistant always being present in order to
properties of the SDQ have been examined in various ensure independent and confidential responding and to
studies, the results of which can be summarised as fol- provide assistance when necessary. Parent participants
lows. To begin with, factor analysis of the parent, teacher, received the questionnaires via their child, completed
and self-report versions of the SDQ has generally indi- them at home, and returned materials in a sealed enve-
cated that the questionnaire consists of five factors that lope. Eventually, complete data sets of 562 children and
correspond with the hypothesised domains of psy- adolescents (254 boys and 308 girls) were available.
chopathology and personal strengths [12, 29]. Further- Mean age of the children was 12.3 years (SD = 1.0, range
more, the internal consistency and test-retest stability of 9–15 years). Percentages of participants from a low, mid-
the SDQ can be defined as satisfactory [12], although it dle, or high socioeconomic background (based on edu-
should be noted that the latter aspect of reliability has cational levels of parents) were 21.2 %, 35.9 %, and
been examined in studies suffering from limitations 42.9 %, respectively.
such as a small sample size [11] and a too long test-retest A randomly selected subsample of 142 children and
interval [12]. Correlations among parent, teacher, and adolescents and their parents was mailed two months af-
self-report SDQ scores are moderate but nevertheless ter the initial assessment with the request to complete
compare favourably to cross-informant correlations as the SDQ (independently) for a second time. Ninety-one
obtained with other psychopathology measures [10, 12, children and their parents (64.1 %; 36 boys and 55 girls,
14]. Finally, evidence has been obtained for the validity mean age = 12.2 years, SD = 0.8, range 10–14 years) re-
of the SDQ. More specifically, SDQ scores correlate sub- sponded to this mailing. These data were used for inves-
stantially with other indexes of psychopathology such as tigating the test-retest stability of the SDQ.
the Rutter and Achenbach questionnaires [10, 16, 20, 21].
Moreover, the SDQ discriminates well between children
with and without psychopathological symptoms [11, 14,
20, 24] and there is evidence to indicate that it can be em-