This paper analyses the design of the EU Profiler, the first truly cross–national VAA. We assess the convergent validity and scaling reliability of the low–dimensional models that are used to represent differences between parties and users. Convergent validity of the party positions in the EU Profiler is moderate to high, but scaling reliability is low for most of the issue dimensions included. We examine whether these problems are related to the EU Profiler's cross–national nature. The EU Profiler integrates the positions of parties from all over Europe into one pan–European model, even though students of European politics emphasise that there are structural differences between party competition in Central and Eastern Europe and Western Europe. We find that the EU Profiler performs better in terms of scaling reliability in Western European party systems than in Central and Eastern European party systems. In addition, there are substantive differences between individual countries.