Variance modeling for intensive longitudinal data
Donald Hedeker
Innovative methods of data collection often produce large numbers of repeated measurements for each individual. Variously known as ecological momentary assessments (EMA), experience sampling method (ESM), and daily diary (DD), these methods have been developed to record the momentary events and experiences of subjects in daily life. Data produced by these methods are commonly referred to as intensive longitudinal data.
This full-day workshop will focus on analysis of intensive longitudinal data using mixed models, also known as multilevel or hierarchical linear models. In particular, extensions to model the variances in mixed models will be the focus. In the standard mixed model, the error variance and the variance of the random effects are assumed to be constant across individuals. For intensive longitudinal data, it becomes practical to allow those variances to vary randomly across individuals, as well as to depend on other covariates including time itself. Besides making the models more realistic, additional substantive insights can be gleaned by modeling both means and variances.
Methods will be illustrated using software, with MixRegLS, SAS, and Stata examples and syntax. Familiarity with reading in data and performing basic statistical analyses in either SAS or Stata is recommended. The freeware MixRegLS program and reading materials can be downloaded from Don Hedeker's website. Stata users will want to install the runmixregls program. Additionally, use of MixRegLS via R will be described.
The workshop will take place in Gentry 325. Registration is free for UCONN faculty, students, and postdoctoral fellows, but registration is limited. Please register by clicking on the following link: Hedeker .
Donald Hedeker
University of Chicago
Bio: Donald Hedeker’s main area of expertise is in the development and use of advanced statistical methods for clustered and longitudinal data, with particular emphasis on mixed-effects models. He is the primary author of four freeware computer programs for mixed-effects analysis: MIXREG for normal-theory models, MIXOR for dichotomous and ordinal outcomes, MIXNO for nominal outcomes, and MIXPREG for counts. In 2000, Hedeker was named a fellow of the American Statistical Association, the highest honor in his field. He serves as an associate editor for Statistics in Medicine and the Journal of Statistical Software. He has been the principal investigator (PI), co-PI, or co-investigator on many research grants funded by the National Institutes of Health (NIH) and the Centers for Disease Control and Prevention. Hedeker earned a PhD in quantitative psychology and a BA in economics from the University of Chicago.
For more information, contact: Betsy at betsy@uconn.edu