Workshop 1:

Continuous Time Models for Intensive Longitudinal Data

Charles Driver

Max Planck Institute for Human Development


Abstract

Continuous time dynamic models allow for flexible and theoretically plausible accounts of the generative process underlying longitudinal data. Latent growth curves and autoregressive panel models arise as specific cases within the more general framework. The “ctsem” package for R is designed to allow for both frequentist and Bayesian fitting of such models, applied to single or multiple subject longitudinal data, with few or many observation occasions, and potentially very different scales of observation timing. Individual differences in some or all parameters may be handled via observed covariates and or random effects. Limited forms of non-stationarity (estimated starting points) are incorporated in most models, while more complex forms (e.g. time varying effects or variances) can be specified but involve alternative estimation approaches. The workshop will provide an introduction to the theory and modelling, then work through a number of examples rising in complexity—linear growth curves, time dependent input effects (such as interventions), multiple interacting processes, individual differences, and non-linear models. Along the way we'll see how to generate data from such models, plot various expectations, and interpret the parameters.

 

The workshops will take place on Sunday, 15.09.2019 from 9:00 to 17:00. Registration for the workshops can be completed via ConfTool.