Sex differences and gender-invariance of mother-reported childhood problem behavior

Sophie van der Sluis, Tinca J C Polderman, Michael C Neale, Frank C Verhulst, Danielle Posthuma, Gwen C Dieleman

Research output: Contribution to JournalArticleAcademicpeer-review

5 Downloads (Pure)


Prevalence and severity of childhood behavioral problems differ between boys and girls, and in psychiatry, testing for gender differences is common practice. Population-based studies show that many psychopathology scales are (partially) Measurement Invariance (MI) with respect to gender, i.e. are unbiased. It is, however, unclear whether these studies generalize towards clinical samples. In a psychiatric outpatient sample, we tested whether the Child Behavior Checklist 6-18 (CBCL) is unbiased with respect to gender. We compared mean scores across gender of all syndrome scales of the CBCL in 3271 patients (63.3% boys) aged 6-18. Second, we tested for MI on both the syndrome scale and the item-level using a stepwise modeling procedure. Six of the eight CBCL syndrome scales included one or more gender-biased items (12.6% of all items), resulting in slight over- or under-estimation of the absolute gender difference in mean scores. Two scales, Somatic Complaints and Rule-breaking Behavior, contained no biased items. The CBCL is a valid instrument to measure gender differences in problem behavior in children and adolescents from a clinical sample; while various gender-biased items were identified, the resulting bias was generally clinically irrelevant, and sufficient items per subscale remained after exclusion of biased items. Copyright © 2016 John Wiley & Sons, Ltd.

Original languageEnglish
Article numbere1498
Pages (from-to)1-15
Number of pages15
JournalInternational Journal of Methods in Psychiatric Research
Issue number3
Early online date22 Jan 2016
Publication statusPublished - Sept 2017


Dive into the research topics of 'Sex differences and gender-invariance of mother-reported childhood problem behavior'. Together they form a unique fingerprint.

Cite this