Research into the effects of neighborhood environments on social and behavioral characteristics is threatened by inattention to some methodological obstacles, including the recruitment of a representative sample of resident participants. Although sampling is relatively straightforward, actual recruitment is not. The authors present the recruitment experience of the Twin Cities Walking Study addressing two questions: (a) Are randomly selected participants different from those selected by convenience? and (b) How well does the realized sample match known demographic characteristics of target neighborhoods? Of 716 total participants nested in 36 neighborhoods, 74% were randomly recruited. Socioeconomic status was positively correlated with random recruitment; nonrandom volunteers were more likely to be non-White females of lower socioeconomic status. Multivariate analyses, using propensity scores, show randomly selected and volunteer participants to be similar. The final mixed sample represented the target neighborhoods well. Supplementing a random sample with volunteers may yield a representative sample exchangeable at the group level.