The present study examines poverty talk in Aotearoa/NZ during the 'poverty debate' of 1996. Social psychology's attribution studies on lay peoples explanations for poverty are reviewed. Methodological limitations with these studies are identified and a discursive approach to the topic advocated. Textual data from the media and two Barnardos Care Centre focus groups are analyzed using methodology developed by Potter and Wetherell (1987), Edwards and Potter (1992), and Wetherell (1998). Four interpretative repertoires from the media text are identified: 'social security as destructive'; 'dissolute character'; 'resourcefulness'; and 'absolutely impoverished'. A further five interpretative repertoires are identified from the focus group talk: 'financial hardship'; 'real DPB person'; 'catch 22'; 'working the system'; and 'resourcefulness'. Focus group talk focusing on the rhetorical construction, and 'situated practice' aspects of the members' accounts during this period is explored, and the ideological implications of the accounting practices discussed.