Aggregation problems in policy evaluation: an overview

Public policy evaluation consists typically of measurements of various criteria, which are then combined into an overall assessment. This paper deals with a few caveats concerning these types of assessments. Some of these pertain to inhomogeneity of observed units, some to the number of criteria applied and some to the combination methods. It turns out that inconsistencies encountered in the evaluation are closely related to those discussed in the theory of voting. Thought experiments are suggested as ways of reducing the significance of evaluation pitfalls.