This paper is motivated by a distinctive appreciation of the difficulties posed by quantitative observational inquiry into complex social and economic systems. It develops ordinary and piecewise indices of joint and incremental informational association that enable robust approaches to a common problem in social inquiry: grappling with associations between a quantity of interest and two distinct sets of co-variates taking values over large numbers of individuals. The distinct analytical usefulness of these indices is illustrated with their application to inquiry into the systemic economic effects of patterns of discrimination by social identity in the U.S. economy.