Towards an understanding of the shaping of research outcomes by contextual issues: reflections on the contributions of the ReMath project