A method for analyzing selection in hierarchically structured populations

Individual fitness depends on the particular ecological, genetic, and social contexts in which organisms are found. Variation in individual context among subunits of a population thus raises interesting questions about selection in nature but also complicates its study. We present a method for analyzing phenotypic selection in hierarchically structured populations. Applying this method, called contextual analysis, to the study of selection allows explicit answers to two frequently controversial questions. First, must group membership be taken into account in explaining differences in individual fitness? Second, what particular group properties are associated with observable group-level effects? Contextual analysis is a generalization to structured populations of the "selection gradient" method developed by Lande and Arnold (1983). The aim of the gradient method is to distinguish characters that have a causal relationship with fitness from others that do not but are still subject to selection as a result of their phenotypic correlations with other traits. This distinction is made using the partial regression coefficients of fitness on the set of characters being investigated. Contextual analysis extends this method to include group as well as individual characters in a multiple-regression analysis of fitness. The partial regression coefficients associated with the group characters estimate the ability of specific group properties to account for variance in individual fitness in excess of what can be accounted for by the phenotypes of the group members themselves. In this way, contextual analysis permits independent group-level effects on fitness to be distinguished from changes in group characters solely as a result of their correlations with individual-level traits. While applicable to a wide range of problems, contextual analysis makes two particular contributions to the study of group selection: it clarifies the much-debated concept of "emergence"; and it provides a powerful analytic tool for studying group selection in nature.