Chapter 5 shows how the methods introduced in the preceding chapters can be used to gain novel substantive and theoretical insights. We show how RIO can be used to identify multiple storylines implied by a single regression model by examining cases (or sets of cases) that contribute to the regression model in otherwise unseen ways. We illustrate RIO’s substantive benefits through empirical analyses of (1) the effects of regional integration on inequality, (2) the social determinants of health, and (3) the correlates of dog ownership.

Survey methods that randomly sample respondents from populations abstract their subjects out of the settings where social phenomena form and develop. By measuring the egocentric networks that surround respondents, surveys can re-incorporate these interpersonal contexts. This chapter reviews approaches to egocentric measurement implemented within the U.S. General Social Survey (GSS). Among these are global items that obtain direct reports about network properties (e.g. size, composition), short sets (aggregated relational data, position generators) that allow estimation of certain network properties, and longer name generator instruments that obtain more granular data on the individual contacts (“alters”) and relationships within a respondent’s egocentric network. The review gives particular attention to the “important matters” name generator for measuring “core” networks, first administered in the 1985 GSS. It covers that instrument’s origins and subsequent use in both substantive and methodological research. Substantive studies show how networks vary by (e.g.) age, socioeconomic standing, gender and residential setting, and offer suggestive evidence about how they shape outcomes including well-being, political activity, and sociopolitical attitudes. Methodological studies reveal that the important matters name generator can be sensitive to several aspects of survey settings, and call for care in its administration.