Evaluating business support measures: The Bayesian Network approach

Abstract

Traditional methods to the evaluation of business support consider the firm as a ‘black box’ and the main interest is to see to what extent a policy has produced the intended effects. The causal mechanisms that explain how certain effects are generated are not discovered. In this article we show the applicability of Bayesian Network Analysis in combination with theory-based evaluation as a new mixed-method approach to reveal the mechanisms, both expected and unknown, which explain the changes in a firm’s behaviour and economic performance due to public support. By combining graphical map analysis with statistical analysis, Bayesian Network Analysis reveals the interdependences between different drivers of change in firms so as to either confirm, deny or better qualify the theory of change of the policy.