|
Abstract:
|
Ecological Momentary Assessment (EMA) studies provide intensively measured longitudinal data with large numbers of observations per unit, and thus provide richer information in understanding people's thoughts, emotions and behaviors compared to traditional studies. While mixed effects regression models (MRM) make it possible to reduce the within unit correlation by introducing random effects, it lacks the flexibility in modeling the within unit variability. We will outline a mixed effect location scale model (MLS) with multiple location and scale random effects that will allow subject as well as wave level heterogeneity in terms of both the mean and within variance of the repeated measurements. Model parameters can be estimated through Bayesian MCMC sampling. Finally we will show results from a simulation study as well as an application to adolescent smoking data. We found that smoking is associated with more consistent positive mood, and that positive mood tends to be more consistent over time.
|
