“…Full descriptions regarding how these samples were constructed in [2002][2003][2004] have been published (Dye, Ducharme, Johnson, Knudsen, & Roman, 2009;Knudsen, Ducharme, & Roman, 2007). Briefly, the strategy relied on random sampling of U.S. counties from 10 population-based strata, identification of all treatment organizations in these counties using published directories, random sampling of organizations, and telephone screening.…”