[This chapter concerns the random survey of the Lower East Side of New York City. The particular methods discussed in this chapter were the results of my contact with the Bureau of Applied Social Research at Columbia, which made it possible for me to profit from the highly professional sampling method used by Cloward and Ohlin in the Mobilization for Youth project (1960). None of the many studies of speech communities that followed used exactly the method developed here, but a good number have been constructed on the same general principles. Chapter 15 provides a summary of the approach taken by 37 sociolinguistic studies of large cities in the years 1966–2006. Some improved on the method for sampling the community outlined here, particularly in the construction of a stratified random sample in which the groups of interest are equally represented, rather than proportional to their numbers in the population. For various reasons, others did not follow the general principle that each member of the community (or sub-group) should have an equal opportunity of entering the sample. In the most different method, the investigator begins with introductions from friends and acquaintances. Such a convenience sample,which follows the researcher's personal networks (sometimes called “friend of a friend” approach), is well justified under some circumstances. Political disturbances may make it impractical to do otherwise, as in Milroy's study of Belfast during the troubles (Milroy and Milroy 1978). The target group may be isolated from contacts by strangers (as in Kroch's (1996) study of the Philadelphia upper class). However, the speakers we contact through personal networks are certain to be more similar to us than the speakers we do not. Though we may reduce the observer effect by interviewing our friends, we will be increasing it by choosing people more similar to ourselves.]