In an urban context people travel between places of residence and work destinations via transportation networks. Transportation studies that involve measurements of distances between residence and work locations tend to use Euclidean distances rather than Network distances. This is due to the historic difficulty in calculating network distances and based on assumptions that differences between Euclidean distance and network distance tend to be constant. This assumption is true only when variation in the network is minor and when self-selection is not present. In this paper we use circuity, the ratio of network to Euclidean distance, as a tool to better understand the choice of residential location relative to work. This is done using two methods of defining origins and destinations in the Twin Cities metropolitan region. The first method of selection is based on actual choice of residence and work locations. The second is based on a randomly selected dataset of origins and destinations in the same region. The findings of the study show circuity measured through randomly selected origins and destinations differ from circuity measured from actual origins and destinations. Workers tend to reside in areas where the circuity is lower, applying intelligence to their location decisions. We posit this because locators wish to achieve the largest residential lot at the shortest commute time. This finding reveals an important issue related to resident choice and location theory and how resident workers tend to locate in an urban context.