Exploring the Relations Among Different Levels of Intraindividual Variability and Longitudinal Change Using a Mixed Effects Location Scale Model

• Activity Number: 335
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Philippe Rast*+
• Companies: University of Victoria
• Keywords: Intensive Measurement Design ; Location Scale Model ; Variance Modeling ; Mixed Effects

Abstract

The investigation of within-person variability has increased in recent years and has proven be a fruitful field in the investigation of cognitive change. At the same time, intensive measurement designs have become more and more important because they allow the estimation of within-person effects. Here, within-person variability is estimated with a mixed effects location scale model. That model allows one to include explanatory variables at the within- and between-person level and it estimates random effects in the location and scale part simultaneously. We used this model in a Bayesian framework to demonstrate its application to three level data. 304 participants have been measured on a reaction time test at four bursts, each one year apart. Each burst contained four sessions of weekly (1 to 2 weeks apart) measurements. We found considerable individual differences among participants in all parameters. The location and scale parameters were not independent whereby the random scale parameters correlated with the random intercept of the location part of the model.