The OECD is extending the participation of low- and middle-income nations in its Programme for International Student Assessment (PISA). To explore how PISA can be made more relevant to these contexts, a pilot study, PISA for Development (PISA-D), was launched. Translating PISA into PISA-D required the development of instruments that had relevance to the new contexts while maintaining comparability across all PISA participants. Drawing on Science and Technology Studies and Callon et al's theory of the three stages of translation of research, and based on detailed empirical data, this paper describes how the technical and the political, and the material and the semiotic, work together to make PISA 'fit' new contexts, while at the same time making the new contexts 'fit' PISA. This paper demonstrates how international comparisons demand profound changes in the ways countries come to know, represent, and act upon their education systems.