Adopting a policy sociology approach, this paper analyses the methods of evaluation used by French high central inspectors since the middle of the eighties on the basis of materials drawn through three complementary qualitative research studies and mobilised here in a synthetic perspective (70 interviews in total, consultation of archives and analyse of various data sets of public and internal documents). It establishes first that inspectors do not publish framing documents on their activities, that they little use their methodological guidebooks and that their new focus on evaluation did not lead to a major change of their methods. Then this situation is explained by various factors like the uncertainties of the current governance of the French education system, the model of evaluation promoted by inspectors, their internal organisation and their very selective use of foreign evidence. It is not interpreted as a lack of expectations but as the consequence of an implicit accountability relationship (Bovens, 2007) which expects mainly from professionals that they comply with fast changing formal requirements and that they control policy outputs rather than its outcomes. This paper thus contributes to the study of the indigenous translations of policy imperatives such as a “management of expectations”.