Psychometric properties of the Children’s Response Styles Questionnaire in a Hong Kong Chinese community sample

BACKGROUND The Children's Response Styles Questionnaire (CRSQ) is a widely-adopted inventory that assesses response styles in youths. It is useful in examining how coping styles (particularly rumination) may relate to depressive vulnerability in youths. Despite its utility, little is known about its applicability in non-Western cultures and CRSQ has not been evaluated using current psychometric methods including item response theory (IRT). The present study assessed the properties using IRT methods in a Chinese youth sample. METHODS Students in Grades 4-6 were recruited from seven public primary schools in Hong Kong, and a total of 581 children (280 boys and 301 girls) between 8 and 14 years of age participated in the study. A Chinese version of CRSQ was administered to them in groups at school after receiving written parental consent as well as students' assent. RESULTS Confirmatory factor analysis revealed a two-factor structure that was comparable to that identified in Western samples, namely, the rumination and distraction/problem-solving subscales. IRT analysis suggested that items varied in levels of item discrimination and item severity, and in precision/usefulness for assessing the underlying latent trait levels. Test information analysis indicated that rumination subscale was more useful than the distraction and problem-solving subscale in assessing the latent trait over a broader range of levels. For gender-based Differential Item Functioning (DIF) analysis, item 1 "When I am sad, I think about how alone I feel" was found to exhibit higher discriminating power for girls than boys. CONCLUSIONS The study presents the first attempt to examine CRSQ item properties using IRT analysis and supports its validity beyond the Western cultures. The factor structure of CRSQ was found to be comparable to the West in our Chinese sample. Differential Item Functioning (DIF) evaluation suggested all but one item in the rumination subscale of the CRSQ apply equally well to both boys and girls.