Researchers attempting to integrate socio-economic data in watershed planning often draw on nationally collected census data. However, there are critical limitations to the usefulness of this type of data for decision makers operating at the watershed scale. In this paper we demonstrate the relevance of spatially referenced socio-economic data collected using mail surveys to random selections of rural landholders. The issue explored was dryland salinity management in two large watersheds in the Murray-Darling Basin of south-eastern Australia. Contrary to the assumptions underlying public policy in Australia, but consistent with the literature on farmer knowledge, comparisons of expert maps and landholder identified salinity sites suggested that landholders in these watersheds had excellent knowledge of the current extent of salinity on their property. Our research also suggested that salinity education was a sound investment by governments. At the same time, the expert maps failed to predict half of the saline affected sites identified by landholders. Accurately mapping the extent of salinity would seem a first step in addressing this nationally significant land degradation issue.