Economics has looked at the decision process of politicians but the decision process of agency staff has primarily been the purview of sociologists. Agencies affect the final form of regulations, they may enforce or ignore regulations that exist, and they provide information to the political process. Policies recommended by economists for non-point source pollution control are seldom implemented by government agencies. This study examined the relationship between preferences for a particular policy and the perceived farmer cost, farmer resistance, efficacy in salinity reduction, fairness, and administrative costs. The latter were included to find whether transaction costs of implementing policies affect preferences and whether this could help explain the existence of current policies. To test this hypothesis, a survey of people working on the salinity issue was conducted. Contrary to what one might expect, perceptions of farmer cost and farmer resistance were not highly correlated. When preference was regressed against farmer cost, farmer resistance and administrative costs, only farmer resistance was significant. When effectiveness and fairness were included as explanatory variables, they were highly significant and the coefficients were quite large. Including perceived effectiveness and fairness greatly improved the explanatory power of the model.