cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
  • Learn how to build custom Python data connectors and further customize JMP’s Data Connector Framework with the Python Data Connector Demo, available now in the JMP Marketplace!
  • See how to create experiments to support product design and ID useful product features. Register for June 12 webinar, 2pm US Eastern Time.

Discussions

Solve problems, and share tips and tricks with other JMP users.
Choose Language Hide Translation Bar
Mickyboy
Level V

Mixed model - Estimate Only Variance Components with 95% C.I.'s

Hi all,

Is it possible to run a mixed model with random effects, and use the estimate only variance component option and request 95% confidence intervals estimates for the variance components, similar to the output you receive when you select unbounded variance components option and request 95% confidence intervals from the output?

 

Thanks,

Mick

1 ACCEPTED SOLUTION

Accepted Solutions
Mickyboy
Level V

Re: Mixed model - Estimate Only Variance Components with 95% C.I.'s

Hi All, 

Thought i had better add a bit more to this, l am running a mixed model with random effects, l have some components with negative variance, so l am using the estimate only variance components, l would like to get 95% confidence intervals in the output if possible, is it possible??? Because l have some negative variance components, l cant use the output from the unbounded variance components option, does anyone have a work around???

 

Thanks,

Mick.

View solution in original post

1 REPLY 1
Mickyboy
Level V

Re: Mixed model - Estimate Only Variance Components with 95% C.I.'s

Hi All, 

Thought i had better add a bit more to this, l am running a mixed model with random effects, l have some components with negative variance, so l am using the estimate only variance components, l would like to get 95% confidence intervals in the output if possible, is it possible??? Because l have some negative variance components, l cant use the output from the unbounded variance components option, does anyone have a work around???

 

Thanks,

Mick.

Recommended Articles