In a traditional paired watershed study, watersheds are selected to be as similar as possible so that conclusions may be drawn about the performance of Best Management Practices. We have extended the paired watershed concept to examine the effectiveness of watershed management programs by adding comparative cri- teria for social characteristics. For four different 8 or 11⁄12 digit hydrologic unit code (HUC) watersheds in the Midwest, we have piloted a systematic method for selecting paired subwatersheds. First, we developed a list of 11 key variables. Next, a factor analysis was conducted to determine the underlying structure of the 11 input variables. Finally, in each of the four watersheds, potential paired subwatersheds (all 14 digit HUCs) were selected using the factors in a cluster analysis. Informal interviews were then held with key informants in each watershed to provide qualitative assessments of criteria that could impact the comparability of the subwater- sheds. This method for selecting paired watersheds should be helpful for other researchers to test the effective- ness of watershed management programs focused on behavior change.