|
Abstract:
|
Models for monotonic change to an asymptote are important in a number of research scenarios. Furthermore, nonlinear mixed-effects versions of these models are required when substantial interindividual differences in the underlying process exist, a common occurrence in the behavioral and medical sciences. Yet, commonly employed parameters for conceptualizing the rate of change toward the asymptote in these models often lack substantive interpretation. Alternative conceptualizations of this change model characteristic, which offer more meaningful interpretation, are presented. Furthermore, it can be especially important to allow for non-normal random-effects and within-individual residual variation that varies across time or the fitted model values when employing the presented models. An illustrative example from a quantitative procedural task is provided.
|
