The development of assessing education systems through the performance of their students is a major phenomenon in recent years. At a national level, several European countries, that have not yet established such assessments, have developed standardised assessments to improve the measure of student performance and used these to infer the quality of schools and teachers. At an international level, The International Association for the Evaluation of Educational Achievement (IEA)1 and Programme for International Student Assessment (PISA) results have become an essential tool for comparing education systems. Their results have launched significant reform movements in some European countries. In the frame of New Public Management, this development marks a shift from steering by inputs to steering by outputs. It also responds to a widespread concern to monitor and improve the quality of educational services in Europe, in line with the Lisbon 2010 objectives. While these external assessments are increasingly at the core of education policies, educational research has tackled questions about their usefulness, limitations and risks. Indeed, output assessments offer a set of information to establish a diagnosis of educational systems, identifying their strengths and weaknesses. Many debates, however, have stressed that the information may be incomplete, reflect only a part of learning, suppress socio-cultural diversity and suffer from some methodological bias. Finally, if output assessments become education's targets, or if they are used to legitimise reforms adopted for other reasons, the data gradually loses its informational value. Allowing researchers to construct and analyse assessment data, according to scientific rules, therefore, is a current challenge in order to ensure their effectiveness in improving educational practices and policies.