Analysis of sex sequences by means of generalized linear mixed models

Adaptive strategies of sex allocation functioning to increase fitness, including strategic allocation of sex in relation to birth order and sex composition of the progeny, have frequently been explored, but the development of a statistical framework for these analyses has lagged behind. In this paper, we contribute to filling this gap by devising a method for analyzing sex sequences based on a proper parameterization of generalized linear mixed models (GLMMs). The method is highly flexible and can be easily extended to investigate sequences of traits or events or other longitudinal data. As a worked example, we focus on the analysis of sex sequences of offspring produced by females in a single reproductive event and develop a method that allows analyzing simultaneously sequences of different length (e.g., clutches of different size) and sequences with missing data, as may frequently happen in "real-world" data sets. Different patterns of allocation among traits or events along the sequence were investigated, and a real data base of sex sequences of eclectus parrots (Eclectus roratus) fledglings produced by different females was analyzed. A tutorial for running the analyses with the R or the SAS software is provided in the Electronic Supplementary Material.