This paper focuses on a neglected area of school policy and practice: metacognition. As education becomes increasingly evidence-informed policy makers, school leaders and teachers are becoming increasingly research literate and have ready access to an ever-growing range of evidence about 'what works' in schools. Influential sources of evidence, such as the Education Endowment Foundation's Teaching and Learning Toolkit, often indicate that teaching metacognition in schools can have a very positive effect on pupils' outcomes. In this paper, we examine over 50 studies to ascertain the effect of teaching metacognition in schools on pupils' outcomes and their wellbeing. Following our review it is clear that there is strong evidence indicating the when metacognition is effectively taught in schools then there is a very positive effect on pupil outcomes; there is less evidence about the relationship between teaching metacognition and pupil wellbeing, but the evidence which does exist is also very positive. Having identified that teaching metacognition can help improve pupil outcomes in schools, we then pose questions about the English government's attitudes towards evidence-based practice. We ask why the government adopts some policies and strategies which have an international evidence base, while not adopting other policies or strategies which have at least an equally strong evidence base. This paper concludes by suggesting how policies and practices can be improved in schools, Initial Teacher Education establishments and at the level of national policy.