Objectifying Contextual Effects. The Use of Snowball Sampling in Political Sociology

The statistical objectification of the contextual dimension of human behaviors, namely the part played by the interactions that individuals have with the people that they encounter in the places they frequent, often seems to be methodological gauntletly challenging. Among the different strategies that propose to achieve this objective, snowball sampling allows for isolating personal networks directly. This article interrogates the conditions of implementation of snowball sampling, as well as how the statistical materials that it produces are used in contextual analysis. I demonstrate why and how multilevel modeling is particularly relevant for solving the issue of the internal validity of snowball samples while taking advantage of the hierarchical structure of data to isolate the influence of contextual variables on behaviors. Referring to my doctoral research and using the specific example of the identification of individual and contextual factors of political discussions, I also show how data obtained through snowball sampling can be mobilized and modelized.