Trip chaining is an important aspect of travel behavior, yet is less well understood than direct trips generated. Using a negative binomial regression model to fit data from the 2009 National Household Travel Survey (NHTS) and the 2010 US Census, this study focuses on the association between the complexity of commute tours (i.e. the number of non-work intervening stops) and the characteristics of commuters, households, and their neighborhoods and regions. The results from the 51 largest metropolitan areas confirm the saliency of household responsibilities, gender, flexibility of work schedule, and household auto ownership, but do not show any strong effect of socio-economic status, the regional and local built environment, or gasoline price. The findings of this study demonstrate the need for travel demand modeling to take into account the effect of trip chaining.