Descriptions of educational inequality, or 'gap talk', require principles of categorisation that divide humans into groups between which a 'gap' can exist. The principles used in education equity policies to define groups affect the nature of the educational interventions that these policies propose. Drawing on critical discourse analysis of education policy documents and political speeches, I argue that in New Zealand the dominant categorising principle is indigenous/Maori ethnicity, whereas in England it is socio-economic disadvantage. These two principles result in two very different types of policy interventions, both of which address only part of the problem. Furthermore, both principles are simplistic and one-dimensional, making it easier to blame 'failing schools' for educational inequality, whereas use of more nuanced principles would provide a better basis for determining the kinds of differences that 'effective schools' can reasonably be expected to make. While previous analyses have suggested that recognition-orientated policies can shift attention away from broader social policies that influence educational disparities, I argue here that this can also be the case for distribution-orientated policies.