Recent interest in the measurement of social capital has generated new suggestions for using survey data to measure basic indicators of social network formation. This paper considers the (2-mode) social networks created by people's involvement in associations. Information on respondents' participation in associations is collected often as an indicator of civic participation however it is also, potentially, a source of information about associational activity among the population or community being surveyed. I build from the work of McPherson (McPherson 1982) who showed how respondent reports on the size of the associations they are involved with, in addition to the fact of their involvement, yields the information necessary to estimate key parameters of network formation. McPherson theorised collective network structures as 2-mode networks, or 'hypernetworks'. I conceptualise these collective entities as systems of association and present them in closely related set-theoretic visualisations, rather than the graph-theoretic visualisations used by SNA. I re-theorise the relationships between participation rates, association size and system parameters to develop indicators of associational coverage overlap and association overlap.