Abstract
Purpose. The purpose of this study was to explain a shift in policy outcomes as it relates to structural changes in policy networks. It looks at a major, controversial issue in the history of California water policy at three different periods to determine what can be learned about the decision and predecision elements of the policy process. Using network concepts and analytic procedures in the empirical diagramming of these water policy networks, the study proposes a typology of networks classified according to their elitist or pluralist elements. Theoretical framework. The study integrates three theories relevant to the distribution of power and its impact on the policy process. Two theories of governance (elitism and pluralism) are presented and studies supporting each of these theories are examined. Network theory is then utilized to map the distribution of power within selected policy networks and explain the policy process. Methodology. This is a comparative, longitudinal case study incorporating a mixed design approach. It utilizes a dominantly qualitative approach (multiple case study) supplemented and enhanced by quantitative analysis. Findings. This study expands on a typology first given by Marsh and Rhodes (1998) and defines five characteristics that are useful in distinguishing network types. The study's findings showed definitive changes in both power (as measured by centrality) and actor classification over a span of ninety-three years. Power was focused in the hands of a few elites in the first time period, was indeterminate or diffused during the second time period, and became slightly more focused in the third period of time. Conclusions and recommendations. The study offers an original contribution in the study of policy networks as they change over time. The study's results showed definite dimensional changes in network structure leading to the conclusion that network characteristics can be quantified and this quantification leads to a richer understanding of network processes. Future studies could expand on this methodology to include macrolevel variables focusing on the wider environment.