Jump to content


Photo

Testing diagonal covariance matrix


  • Please log in to reply
2 replies to this topic

#1 Davidq8

Davidq8

    Advanced Member

  • Members
  • PipPipPip
  • 57 posts

Posted 01 October 2022 - 02:18 AM

I have a question regarding testing diagonal covariance matrix (Full and partial block) in a model. Should we do it after adding the covariate effect into the final model, or before the stepwise process (base model)?

Thanks!

 

 


Edited by Davidq8, 01 October 2022 - 02:28 AM.


#2 smouksassi1

smouksassi1

    Advanced Member

  • Members
  • PipPipPip
  • 231 posts
  • LocationMontreal

Posted 04 October 2022 - 07:56 AM

there is two schools of thought about this, newer engines like QRPEM would be happy with a full omega and then you can reduce it if you see fit

 

adding a covariate can reduce the off diagonals ( explain omega correlations) and also the diagonals

 

see the figure 2  diagmram here (full omega before doing covariate selection)

https://pubmed.ncbi....h.gov/23836283/

 

I typically  avoid doing scm (stepwise) there is tons of literature showing that it has major drawbacks:

https://ard.bmj.com/.../77/Suppl_2/5.1

 

you should look at full model  that are plausible and useful to answer key clinical questions


Edited by smouksassi1, 04 October 2022 - 07:59 AM.


#3 Davidq8

Davidq8

    Advanced Member

  • Members
  • PipPipPip
  • 57 posts

Posted 06 October 2022 - 03:40 PM

Your detailed answer is really helpful, Thank you Sameer. 






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users