Developing and refining social norms measures for the evaluation of a school-based intervention to reduce dating and relationship violence among adolescents in England

Introduction: Dating and relationship violence (DRV) is widespread in England. DRV is associated with increased prevalence of sexual risk behaviour and poor mental health. Interventions often aim to shift harmful social norms underpinning DRV, but lack of valid, reliable measures is a barrier. Methods: I conducted a systematic review of DRV social-norms measures. I developed three brief measures of social norms concerning DRV and gender. I refined these using cognitive testing and assessed the reliability and validity of resulting measures using student surveys in five secondary schools. I analysed qualitative data from students, staff and parents and carers in ten secondary schools to explore how social norms are implicated in DRV in England and inform further measure refinement. Results: Most of the 40 social-norms measures identified in the review were associated with DRV outcomes. Other evidence of reliability and validity was mixed and no measure was shared across studies. In cognitive testing of social-norms measures, answerability was improved where items assessed norms salient and publicly manifest among a cohesive, influential reference group. Refined measures were tested among 1,426 students (82.5% response rate). While floor-effects indicate limited sensitivity to low-to-moderate levels of the assessed constructs, all three measures were reliable and valid. Qualitative interviews suggest that DRV is sustained directly by norms tolerating controlling behaviours and inhibiting disclosure of victimisation, and indirectly via sexist norms that subjugate girls to boys and facilitate gender-based harassment and abuse. Discussion: My findings support the reliability and validity of the three tested measures of social norms, which can be incorporated into evaluations. Research to assess the measures’ cross-cultural validity would contribute to improving comparability of norms across contexts. Further research should seek to develop longer-form versions with increased sensitivity for use in epidemiological research. My findings support recommendations for improving social norms measurement in DRV research.