Attribution og contribution: kausalitet, forandringsteori og mekanismer i evaluering

English AbstractThis dissertation examines the construction of knowledge in three currents of the evaluation field: experiments and evidence-based meta-studies, realistic evaluation, and theories of change. Central to evaluation is the question of causality: which effects can be attributed to which causes? In order to answer the question of attribution, however, an understanding and conceptualisation of the programme being evaluated and its contextual boundaries is required.Evaluation models based on theories of variance and models based on theories of process give very different answers to these questions. The experimental tradition addresses the problem of attribution by isolating the programme from contextual factors by a counterfactual analysis. On the other hand, models based on programme theory address context by making contextual conditions part of the explanation, asserting that only context-dependent explanation is possible. In complex contexts a counterfactual analysis may neither be possible nor relevant. The question of attribution may therefore best be addressed by its transformation into a question of contribution: did the programme make a difference, all things c