Why does "multcompare" output different results in R2014b when data is processed with "fitrm" versus "anova1"?

1 view (last 30 days)
The results obtained from my repeated measures analysis are quite unexpected. For the attached data, I am trying to compare 11 time points. Standard one-way ANOVA ("anova1") says the 9th time point is significantly different from many other time points. However, the "multcompare" for repeated measures ANOVA says that the 9th time point is NOT different from any of other time points (p-value > 0.05). This is highly unexpected since one would hope to get higher statistical power with the repeated measures ANOVA approach. How is "fitrm" followed by "multcompare" different than calling "anova1" followed by "multcompare"?

Accepted Answer

MathWorks Support Team
MathWorks Support Team on 6 Mar 2015
The "fitrm" function models the covariance between observations across "Time" in the attached example whereas the "anova1" function does not. Additionally, the "multcompare" method of a RepeatedMeasuresModel accounts for this covariance whereas "anova1" followed by "multcompare" does not.  This is expected behavior and is responsible for the different results.

More Answers (0)

Tags

No tags entered yet.

Products


Release

R2014b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!