cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
JMP is taking Discovery online, April 16 and 18. Register today and join us for interactive sessions featuring popular presentation topics, networking, and discussions with the experts.
Choose Language Hide Translation Bar

How to measure the effect size of fixed effects in a mixed model including random factors ?

Hi,

I am analyzing a model (Personality: Standards Least Squares, Method REML) with 3 Fixed Effects and 4 random Effects.

I can see on the report the contribution of the variance of each random effect, but I cannot find where can I calculate the effect size of my fixed Effects.

Any help, will be greatly appreciated.

Cheers

Jennifer.

1 ACCEPTED SOLUTION

Accepted Solutions
Kevin_Anderson
Level VI

Re: How to measure the effect size of fixed effects in a mixed model including random factors ?

Hi, Jennifer!

You don't mention what version of JMP you're using for your analysis.  If you're using JMP11Pro, there's a Mixed Model personality in the Fit Model platform that clearly outputs a plethora of Fixed Effects reports.  Don't despair if you aren't using JMP11Pro...although an upgrade couldn't hurt... 

Random and Mixed effects models are sometimes referred to as Variance Components models.They have a long history, stretching back to at least 1861, with voluminous literature, especially recently.  In particular, I would point you towards Milliken and Johnson's Analysis of Messy Data Volume I, which has an entire section on the Analysis of Fixed Part of the Mixed Model.  The same George Milliken wrote a paper with Annie Dudley and John Sall for SUGI25 in 2000 called "Mixed Models Analysis Using JMP® Software 4.0", so JMP has righteously supported Mixed Model analysis for quite a while!

The primary interest in Variance Component models is in the estimation of the variances of the random effects.  That's why the Standard Least Squares report seems to concentrate on them.

Find the Least Squares Means table.  It will give you estimates of the effect sizes.

View solution in original post

2 REPLIES 2
Kevin_Anderson
Level VI

Re: How to measure the effect size of fixed effects in a mixed model including random factors ?

Hi, Jennifer!

You don't mention what version of JMP you're using for your analysis.  If you're using JMP11Pro, there's a Mixed Model personality in the Fit Model platform that clearly outputs a plethora of Fixed Effects reports.  Don't despair if you aren't using JMP11Pro...although an upgrade couldn't hurt... 

Random and Mixed effects models are sometimes referred to as Variance Components models.They have a long history, stretching back to at least 1861, with voluminous literature, especially recently.  In particular, I would point you towards Milliken and Johnson's Analysis of Messy Data Volume I, which has an entire section on the Analysis of Fixed Part of the Mixed Model.  The same George Milliken wrote a paper with Annie Dudley and John Sall for SUGI25 in 2000 called "Mixed Models Analysis Using JMP® Software 4.0", so JMP has righteously supported Mixed Model analysis for quite a while!

The primary interest in Variance Component models is in the estimation of the variances of the random effects.  That's why the Standard Least Squares report seems to concentrate on them.

Find the Least Squares Means table.  It will give you estimates of the effect sizes.

Re: How to measure the effect size of fixed effects in a mixed model including random factors ?

Thanks...