In the Programme for International Student Assessment (PISA) the items are organised in small clusters relating to the same stimulus material (called 'units'). Homogeneity analysis (HA) is used to develop a detailed description of the relationship between all the items in one unit, using the categorical information available in the PISA data. The main findings of the analyses presented are the following: (a) non-respondents are separated from those responding; (b) student patterns across the nominal variables cannot be used to develop a detailed insight into how students' thinking and knowledge is organised beyond the fact that good students perform well on most items; and (c) the profiles for countries are described by the same dimensions, indicating that the relative success of the countries can to a certain degree be explained differentially by these dimensions. From these findings implications for future assessments are suggested.