Goals of evaluation and types of evidence

All stakeholders are urged to pay more attention to the quality of evidence used and produced during the evaluation process in order to select appropriate evaluation methods. A `theory of evidence for evaluation' is needed to better address this issue. This article discusses the relationships between the three main goals of evaluation (to learn, measure and understand) and the various types of evidence (evidence of presence, of difference-making, of mechanism) which are produced and/or used in the evaluation process. It argues for the need to clearly distinguish between this approach and that of levels of evidence, which is linked to data collection and processing methods (e. g. single case observations, difference methods, randomized controlled trials.). The analysis is illustrated by examples in the field of agro-environmental policymaking and farm advisory services.