Purpose This study examines how digital pedagogy integration and recognition of prior learning (RPL) frameworks influence educators' ability to deliver innovative curricula in Bangladeshi TVET institutions. It further investigates the mediating role of educator instructional self-efficacy and the moderating effect of institutional resource capacity. Design/methodology/approach A mixed-method qualitative approach was employed, combining a systematic literature review of 20 peer-reviewed TVET articles with semi-structured interviews conducted with twelve institutional stakeholders (Director of TVET Institute, Head of Curriculum Development, Industry Liaison Manager, Senior TVET Instructor, Prior Learning Recognition Coordinator and ICT Training Manager). Thematic analysis and NVivo-assisted coding were used to identify core themes, while Python-based topic modelling and co-occurrence heatmap visualisation were applied to the SLR dataset to map emerging research agendas. Findings The results show that digital pedagogy integration (virtual labs, blended-learning readiness and multimedia utilisation) and RPL frameworks competency mapping and e-portfolio assessment strongly support improvements in perceived curriculum innovativeness. These relationships are mediated by educator instructional self-efficacy, which is strengthened through professional development, peer mentoring and reflective practice. Institutional resource capacity positively moderates these links by providing technical infrastructure, professional support and industry collaboration. Originality/value This research extends technology acceptance model (TAM) and social cognitive theory by demonstrating that perceived curriculum innovativeness results from the alignment of institutional innovation frameworks, educator agency and organisational support. The use of Python-based text mining together with qualitative analysis offers a novel moderated-mediation model that supports both theoretical refinement and practical guidance for enhancing curriculum innovation in resource-constrained TVET settings.
Hasan et al. (Thu,) studied this question.