Online Program

Saturday, February 23
PCE2 Mixed Model Power Analysis by Example: Using Free Web-Based Power Software Sat, Feb 23, 1:45 PM - 3:30 PM

Relevance to Conference Goals (302596)

The session provides an interactive demonstration of power and sample size calculations for longitudinal and multilevel studies. The session will educate attendees about a simple method to calculate power for a wide variety of common mixed model analyses. The power and sample size methods will be useful for statisticians working with longitudinal and multilevel data. The session is targeted at applied statisticians working in academic, industry, or government settings.

Effective communication between statisticians and scientists is required to construct a valid power analysis. Attendees will learn a series of simple questions to ask collaborators. The questions guide the identification of the study design, the classification of hypotheses, and the selection of a covariance model. By expressing complex covariance structures in terms of simpler pieces, researchers will be able to perform valid power analyses for complex longitudinal and multilevel designs.

Session attendees will receive detailed lecture notes describing the sample size analyses and step-by-step instructions for using the GLIMMPSE software. In addition, attendees will be given access to an archive of completed power and sample size analyses. With access to these resources, attendees will be able to implement the methods in their daily practice and train colleagues. Since the software is accessible through a standard browser, statisticians and their colleagues can use GLIMMPSE without installation or licensing hassles. Thus, statisticians will be able to immediately incorporate power and sample size methods for mixed models into their work.