Psychometric validation of the teachers’ responses to bullying questionnaire (TRBQ) among Spanish adolescents

Abstract Background Adolescents’ perceptions of teachers’ responses to bullying are relevant to understanding how adult intervention is experienced in school settings. Although the Teachers’ Responses to Bullying Questionnaire (TRBQ) has been validated in several countries, previous studies have reported different factorial solutions. This study examined the psychometric properties of the TRBQ among Spanish adolescents. Methods A psychometric validation study was conducted with 1,279 adolescents aged 12–16 years enrolled in secondary schools in Spain. A cross-validation approach was used, with exploratory factor analysis (EFA) in one subsample ( n = 639) and confirmatory analyses in a second subsample (initial n = 640; analytical n = 624). Competing factorial models, internal consistency, convergent and discriminant validity, measurement invariance across sex and school grade, missing-data sensitivity, and two-week test–retest reliability were examined. Results EFA showed overlap between Non-intervention and Mediation, and Item 3 was excluded because of inadequate psychometric performance. Among the models tested, the 14-item correlated five-factor model showed the best fit and was retained as the preferred representation, although the four-factor and second-order models also showed acceptable fit. The preferred model showed good fit, χ²robust (67) = 139.447, CFIrobust = 0.979, TLIrobust = 0.972, RMSEArobust = 0.052, 90% CI [0.040, 0.065], and SRMR = 0.032. Internal consistency was adequate for the four active-response dimensions, whereas the two-item Non-intervention factor showed borderline reliability and narrower content coverage. Convergent validity was supported, whereas discriminant validity was limited for some factor pairs. Strict invariance was supported across sex and scalar invariance across school grade, whereas strict grade invariance was borderline. Test–retest reliability was moderate (ICC = 0.557–0.668), with greater uncertainty for Non-intervention. Longitudinal analyses provided preliminary evidence of approximate metric invariance, whereas exact scalar invariance was not fully supported. Conclusions The findings provide additional evidence for a promising 14-item, five-factor interpretation of the TRBQ among Spanish secondary-school adolescents, but the factorial solution should not be considered definitive. The substantial owarrantsamong active-response dimensions and the borderline reliability and limited content coverage of Non-intervention warrant cautious interpretation. The TRBQ is best considered a research measure of adolescents’ perceptions of teachers’ responses to bullying. Further evidence of external and predictive validity, sensitivity to change, longitudinal invariance, and independent replication is needed before broader applications can be recommended.