This article explores the potential pitfalls of using surveys on organisational practices, firms’ technological choices, and workplace relations. Using a generic methodology, one can, under certain hypotheses, estimate the degree of truthfulness or consistency of responding executives working in different local units of a single enterprise. Applied to the French REPONSE survey in regard to the enterprise’s general characteristics (such as size, ties to an enterprise group, and shareholder structure), the methodology suggests that (1) executives usually give rather consistent answers to a question handled separately from the others, and (2) that their answers are all the more specific as the questions are simple and fall within the respondent’s sphere of competency. However, responses to questions on social relations and trade-union representation in the firm are less reliable. Allowing for executives’ errors, we are led to substantially revise both the level and the rate of change of a number of variables, particularly the weights of different employees’ unions in firms.