This realist review addresses the question: 'Under what circumstances does enhancing community accountability and empowerment improve education outcomes, particularly for the poor?' Community accountability and empowerment interventions, it has been
argued, improve educational outcomes by improving the quality of educational services and the participation of students and families in education. However, there has been no agreed understanding of what is meant by ‘community accountability’ or ‘community
empowerment’ in relation to education. The range of interventions which, it has been claimed, affect accountability and empowerment, is broad, and evidence of impacts has been mixed.

Search strategies included keyword searches in numerous databases, including IngentaConnect; JSTOR: The Scholarly Journal Archive Icon; ProQuest; UNESCO Information Sources; document searches of relevant websites; keyword and targeted searches using Google Scholar; snowballing of references of included documents and consultation with End User group members. Over 21,000 individual documents were identified. Titles and/or abstracts were considered against inclusion and exclusion criteria. ‘Core’ documents provided information on accountability and empowerment interventions, and provided data on intermediate or final education outcomes. Other documents were included if they provided data relevant to processes of change, mechanisms, or contextual features affecting whether and how interventions ‘worked’ (or did not).

At the beginning of the research, an initial programme theory was developed for the overall class of community accountability and empowerment interventions, using a rough hierarchy-of-outcomes format. Evidence relating to outcomes for education,
empowerment and accountability was aligned with that initial hierarchy of outcomes, and revisions were made to address outcomes not covered by the initial rough theory. Programme mechanisms were identified abductively. Features of context that appeared to
affect the operations and outcomes of interventions were identified through close reading of texts and propositions about context (abstracted to the level of middle-level theory) were drafted. Evidence from a wider selection of texts was aligned against the mechanism and context propositions. A CMOC (Context-Mechanism-Outcome Configurations) table was developed by aligning significant features of context against mechanisms, either on the basis of evidence or on the basis of logic (where evidence was not available). A theoretical model for the relationship between empowerment and accountability was proposed on the basis of the findings. The CMOC table and the empowerment and accountability model constitute the revised theory that is the intended product of a realist review. Ways in which the theory might be used to support practice are described below. That theory remains to be tested and further refined through future research and evaluation. Implications for policy and practice and for future research and evaluation have been identified. While descriptions of community accountability interventions often focus on a simple programme theory, we have found that effective interventions work through a combination of mechanisms and require a combination of strategies. They take into account factors at national, sub-national and local levels and changing circumstances. Interventions must clarify the types of accountability they intend to address (of whom, to whom, for what, within which power relationships, and so on) and be tailored to local contexts if they are to be effective.

Research and evaluation should similarly identify the different mechanisms that are expected to operate and explicitly gather and make available data to better understand them and the contexts within which they work.