Disclosed is a method for detecting a probable error in a set of data relative to a wearer and/or to a frame chosen by the wearer, at least one of these data being used for ordering an ophthalmic lens to be fitted into the frame for the wearer, the set of data including at least a value of a first parameter and a value of a second parameter different from the first parameter, the first and second parameters being relative to the wearer and/or a frame chosen by the wearer, the method including: a) determining a level of consistency of the values of the first and second parameter with each other, based on a predetermined statistical consistency rule linking the first and second parameters; and b) emitting, based on this level of consistency, an alert signal in order to report a probable error.