Originating from the learning sciences community, design-based research (DBR) is attracting interest from many educational researchers including those focused on mathematics. Beyond its research role, it is being seen as a collaborative way to engage teachers in deep professional development leading not only to changes in skill but also in purpose. Previous reviews of the approach, however, have offered only cautious optimism for the capacity of DBR to support widespread or scalable change. This paper will use methods drawn from the "digital humanities and social sciences" to explore patterns relating to regional differences, theoretical underpinnings, practical implementation and methodological choice in recent DBR research in the domain of mathematics education. The findings suggest that much of the work presented as DBR might be better characterized as 'implementation studies' as they contain only limited commitment to theoretical development capable of supporting the scaling of innovation. The exceptions appear to occur in settings with well-developed research capacity which recognizes the need for ontological innovation of theory whilst iteratively and comprehensively exploring the complexities of authentic learning interventions. An example of a rich mathematical research study using DBR is provided to point to the potential for this methodology to achieve its overarching aims more fully.