L’évaluation d’impact territorial en pratique. Réflexion sur les représentations conventionnelles des logiques d’action publique

The territorial impact assessment of programs is part of the evaluation of the effectiveness of public policies and programs in a geographical and institutional territorial perimeter. It is therefore an ambitious activity of knowledge at the service of public action for territorial purposes, before, during and after the action. The evaluation activity may involve counterfactual production processes. It can be satisfied with conventional representations of the logic of public action. Only the latter are analyzed here in depth in terms of feasibility and reliability. The conventions of logics of action, when shared between the actors of impact assessment, are at the heart of the specifications and serve as a framework for the terms of reference of evaluations: they are the language shared by public actors and evaluators. The article refers to the practice of these representations in evaluators' practices only: the emphasis is on simple conventions of ?impact chain? representations, but also, in the case of complex policies, on representations of ?value and impact networks?. Since these are conventions, questions of feasibility and reliability of evaluation results arise immediately: they are dealt with here in a practitioner's spirit. It is shown that a confidence indicator in the results often confuses the evaluation results, but this may increase the acceptability of the territorial impact evaluation results, which, in addition, may encourage the use of evaluation results.