A Bayesian Network approach to developing effective policies for more sustainable rural land use and development is presented which:
• incorporates several different stakeholder viewpoints
• integrates and balances the results from different viewpoints, and thus facilitates compromise land use solutions
• elucidates the main factors affecting stakeholders' decisions, on which policy levers will operate more effectively
• allows differences between "expected change" and "observed change" to be understood as arising from uncertainty and variation about "the average situation"
• demonstrates the crucial roles of beliefs and uncertainties in determining the preferred options of different groups
• emphasises the need to know the beliefs and uncertainties of local land managers
• incorporates public participation and social learning while such necessary data are gathered.
The approach's merits for fostering pro-active policy design and subsidiarity, and for focusing the research needed to underpin rural policy are discussed.