NONLINEAR MIXED‐EFFECTS MODELS FOR REPEATED MEASUREMENTS DATA

This chapter considers the specification of a nonlinear mixed-effects model for repeated measurements data. Under mixed-effects modeling, individual responses follow a similar functional form but with parameters that vary across individuals. Hence, a mixed-effects model contains both fixed and random effects. With repeated measurements data, the authors observe a number of individuals repeatedly under differing experimental circumstances, where the individuals are assumed to be drawn randomly from a specific population. Given a set of repeated measures on an individual or subject, one can identify two sources of variation in the data: random variation among observations associated with a given individual (intraindividual variation) and random variation occurring among individuals (interindividual variation). The chapter explains some special cases of the hierarchical global model. PROC NLMIXED is used to estimate the parameters of growth curves in which both fixed and random effects enter the response function in a nonlinear fashion.