This study adopted a quantitative methodology involving a matched sample of federated and non-federated schools and utilising multilevel modelling techniques to explore the impact of federations on student outcomes. The sample involves a total of 50 school districts and 264 schools. These are grouped into 122 federations; 264 comparator schools were selected to match these. The study has identified 6 broad and sometimes overlapping categories of federations. The findings suggest there is evidence of impact on overall performance, in that, while federation and comparator schools perform similarly at baseline, federation is positively related to performance in the years following federation. There is evidence to suggest that impact is strongest in performance federations. There is no relationship between federation and inspection judgements or differential impact on students from different socioeconomic settings. In conclusion, this article reflects on the findings to discuss the key implications for future research, policy, and practice.