Online Program

Mixed Models for Ecological Momentary Assessment (EMA) Data

*Donald Hedeker, University of Illinois at Chicago 

Keywords: complex variance, intensive longitudinal data, location-scale models

Modern data collection procedures, such as ecological momentary assessments (EMA), experience sampling, and diary methods have been developed to record the momentary events and experiences of subjects in daily life. These procedures yield relatively large numbers of subjects and observations per subject, and data from such designs are sometimes referred to as intensive longitudinal data. Data from EMA studies are inherently multilevel with, for example, (level-1) observations nested within (level-2) subjects. Thus, mixed models (aka multilevel or hierarchical linear models) are increasingly used for EMA data analysis. In this workshop, use of mixed models for analysis of EMA data will be described with specific focus on analyses of data from an ongoing adolescent smoking EMA study. An important issue that will be described is the treatment of occasion-varying covariates, and the decomposition of the within-subjects (WS) and between-subjects (BS) effects of such covariates. Furthermore, because there are so many measurements per subject, models for relating covariates to the WS and BS variance will be described, including mixed location-scale models that include random subject scale parameters. Such random scale parameters allow subjects to vary in terms of their variance, or scale, in addition to the more typical random subject location effects. Such extended mixed models can be used to assess the determinants of inter-individual and between-subjects variation. Examples will be presented which focus on the variation of mood that is associated with smoking, and the degree to which subject characteristics influence the mood variation. SAS syntax, available via, will be provided and described to facilitate use of the models presented in this workshop. Though the focus will be on analysis of EMA data, the methods have applicability to other types of “intensive” longitudinal datasets.