I describe a major research program on primary mathematics in the UK, the Leverhulme Numeracy Research Program, and give a sample of some of the preliminary results. The program combines large-scale longitudinal survey and case studies, quantitative and qualitative data, and observation and intervention studies in order to try to ascertain how and in what way different factors affect pupils' progress. The case studies have focused on factors relating to home cultures, pupil behaviors, curriculum and teaching styles, teacher subject knowledge, and school policies/leadership. The results quoted relate to these areas and also to effects of the implementation of a national reform which occurred during the research program. (Author)