Friday, December 18, 2009

CTS Workshop 12/18/09 Notes

Present: Amanda, Ben, Abby, Kairn, and Rodger

Abby Crocker lead discussion on the article: "A Comparison of the General Linear Mixed Model and Repeated Measures ANOVA Using a Dataset with Multiple Mission Data Points."

In her work, Abby had been looking for an article that answered questions about how, why, and when to use General Linear Mixed Model (GLMM) measures and found a hole in the literature. She is considering writing a tutorial article for clinician researchers. She also wonders if GLMM, which appears to work for any kind of data, is “too good to be true.”

Here were some questions proposed and discussed:
-What are the advantages and disadvantages of GLMM?
-If you have a complete dataset, when does it make sense to use GLMM? When not?
-When is GLMM better than repeated measures ANOVA?

Ways to make a GLMM tutorial more understandable to clinicians may be:
-Use clinical rather than statistical terms in the bulk of the article
-Create visual examples of GLMM
-Use clinically interesting data
-Don’t compare to ANOVA
-Presenting an algorithm for when to use GLMM (Predictors: time, any data type; Outcomes: Any, except maybe nominal; Assumptions: some missing data OK, sample size limitations uncertain, must assert type of relationship such as linear or quadratic)

Next session: Jan 8.
I was called out of the room as the time ended. It is not clear if future topics were discussed.

Kairn Kelley

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.