Open science interventions to improve reproducibility in science: a narrative review from a behaviour change wheel perspective

Le résumé fourni par la source

Abstract Academics, policymakers and other research stakeholders have broadly discussed strategies to improve research reproducibility, with heightened attention given to open science practices. One strand of these discussions adopts a behavioural perspective centred on researchers. This narrative review applied the behaviour change wheel (BCW) framework to analyse 103 interventions evaluated for their effectiveness in improving reproducibility and reproducibility-related practices. We systematically mapped the included interventions across the nine BCW intervention types and seven supporting policy categories, highlighting which mechanisms have been most and least frequently evaluated empirically. We then interpreted these mechanisms using the COM-B behaviour model (capability–opportunity–motivation). The analysis revealed that ‘coercion’ and ‘persuasion’ were the most commonly studied types of intervention, implemented through ‘regulations’ and ‘guidelines’, with active enforcement appearing to be more effective than voluntary adherence. By contrast, ‘incentivisation’, ‘training’ and ‘modelling'—central to many reform agendas—were rarely evaluated empirically, and ‘communication’ and ‘legislation’ seldom appeared as implementation policies. Furthermore, most interventions were journal-led rather than funder- or university-led, indicating the limited involvement of other important stakeholders. Our analysis suggests that the key to supporting reproducible and transparent research lies in a strategic and integrated approach that targets all components of COM-B, with accountability mechanisms, context-sensitive measures and synergistic action involving multiple stakeholders.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.