Investigating integrity

This chapter is the second of three methodological chapters (Chapters 12–14) for this edited collection. In Chapter 12, the authors discussed influence diagrams. Chapter 13 builds on these influence diagrams to explain how one research team parameterised two of their influence diagrams, leading to the creation of a Bayesian network ( BN ), or Bayesian belief network ( BBN ). While the influence diagrams captured a qualitative analysis of the components shaping integrity, the BN now presents a quantitative layer of analysis by endowing the nodes of influence diagrams with states and probabilities. Parameterising influence diagrams was critical to transforming them into a single BN. To perform parameterisation, the research team conducted indirect expert elicitation through a multi-professional focus group. The values distilled from the expert elicitation process led to a baseline, Scenario 1. Two sets of hypothetical modifications were then made on specific network variables within this baseline scenario, leading to Scenarios 2 and 3. Scenario 2 reflected a case where the government used regulation, resulting in changes in probabilities in four nodes. Scenario 3 reflected a case where the government used client leadership, leading to changes in probabilities in eight nodes. Findings suggest that while Scenario 3 initially appeared preferable, appearing to lock in more certainty across a broader range of nodes, Scenario 2 actually brought about better outcomes.