This paper describes the different statistical measures of criminality that have been used in France since the 1970s and the policies for which they have been used. They first bore on the police activity, then on the fear of crime and on the preoccupation with security, then on the victimization of citizens. Finally, since the 2000s, these data have been used as performance measures of the police force this practice being denounced by those concerned for creating many problems. We explain these statistical measures by questioning the link between expertise and publicity.