New evidence of factor structure and measurement invariance of the SDQ across five European nations

Abstract

International audienceThe main purpose of the present study was to analyse the internal structure and to test the measurement invariance of the Strengths and Difficulties Questionnaire (SDQ), self-reported version, in five European countries. The sample consisted of 3012 adolescents aged between 12 and 17 years (M = 14.20; SD = 0.83). The five-factor model (with correlated errors added), and the five-factor model (with correlated errors added) with the reverse-worded items allowed to cross-load on the Prosocial subscale, displayed adequate goodness of-fit indices. Multi-group confirmatory factor analysis showed that the five-factor model (with correlated errors added) had partial strong measurement invariance by countries. A total of 11 of the 25 items were non-invariant across samples. The level of internal consistency of the Total difficulties score was 0.84, ranging between 0.69 and 0.78 for the SDQ subscales. The findings indicate that the SDQ's subscales need to be modified in various ways for screening emotional and behavioural problems in the five European countries that were analysed

Review of the strengths and difficulties questionnaire translated into languages spoken by children and adolescents of refugee background

“…The CFA model may be varied to allow identification of partial measurement invariance by imposing equality constraints on item thresholds, factor loadings or error variances across groups. Significant differences between the constrained and unconstrained models indicate full invariance is not supported (Milfont & Fischer, ; Ortuño‐Sierra et al, ). Finally, scalar invariance – justifying mean score comparisons – is tested by constraining item intercepts to be equal across groups (Milfont & Fischer, ).…”

“…Other inter‐country comparisons of the S‐SDQ, such as made by Stevanovic et al (), also found poor fit for a 5‐factor model, with a 3‐factor model providing a better but non‐invariant fit across five European countries (Germany, Cyprus, England, Sweden, Italy; Essau et al, ), and partial measurement invariance with DIF for 11 items across Spain, England, Ireland, Germany and France (Ortuño‐Sierra et al, ). Ortuño‐Sierra et al () suggested that greater cultural differences may explain greater disparities in factor structure found by Stevanovic et al (), who compared Turkey, Serbia, Bulgaria, Croatia, Inda, Indonesia, and Nigeria.…”

“…ROAMER aims to improve translation of research into practice and identification and treatment of mental disorders, as exemplified in the work of Essau et al () and Ortuño‐Sierra et al (). Collaborative research has particular salience in the context of the current movement of child, adolescent and adult refugees from the Middle East, and other countries, and would be consistent with ESCAPs call for cross‐country cooperation in screening and responding to young refugees' mental health needs (Gadeberg & Norredam, ; Hebebrand et al, ).…”

“…This articles' findings of structural (and hence construct) inequivalence across many countries and languages underscore the need to conduct cognitive interviews across cultures to investigate how respondents understand items, with findings incorporated into a revised, culturally sensitive SDQ. Such a project would create an opportunity for the type of cross-country research collaboration and integration proposed by ROAMER aims to improve translation of research into practice and identification and treatment of mental disorders, as exemplified in the work of Essau et al (2012) and Ortuño-Sierra et al (2015). Collaborative research has particular salience in the context of the current movement of child, adolescent and adult refugees from the Middle East, and other countries, and would be consistent with ESCAPs call for cross-country cooperation in screening and responding to young refugees' mental health needs (Gadeberg & Norredam, 2016;Hebebrand et al, 2016).…”

Identifying Patterns of Emotional and Behavioural Problems in Preschool children : Facilitating Early Detection

“…There appear to be some cultural differences in behaviour, or in how preschool teachers already estimate behaviour at preschool age. According to work conducted by Ortuño-Sierra et al (2015), various modifications of the present SDQ scales could be required if emotional and behavioural problems are to be screened in different European countries. It is difficult to comment upon differences around the world in general as the SDQ has been mostly studied and used in Europe and Australia (Borg et al, 2012;Warnick, Bracken & Kasl, 2008).…”