This extensive report presents the recommendations and findings of a study designed to indicate procedures and instruments for a national study of the incidence of runaway. Reported data are from an urban-suburban area and a rural area in Colorado. The report makes specific recommendations concerning the feasibility of using a national probability sample for the purpose of estimating the incidence of running away. Several smaller studies conducted in several carefully selected localities are recommended for collecting detailed information about runaway episodes, runaways and their families. The report also presents incidence estimates, an analysis of user satisfaction with sources of assistance, descriptions of types of runaways and runaway behavior, and global comparisons between runaways and nonrunaways. The social-psychological theories relating to runaway behavior are examined, and the methodology and results of the pilot study are described in detail. (SJL)