Missing value analysis/ Little's mcar test

Discuss the jamovi platform, possible improvements, etc.

by delvecct » Fri Jun 14, 2019 6:27 pm

How do I use Jamovi to determine whether missing data is mcar? Is there a Little's mcar test somewhere?
delvecct
 
Posts: 1
Joined: Fri Jun 14, 2019 6:21 pm

by jonathon » Sun Jun 16, 2019 11:45 pm

hi,

i don't know what mcar is, and i don't think jamovi provides it - but if you figure out how to do it in R, you'll be able to run it in jamovi with the Rj editor.

https://blog.jamovi.org/2018/07/30/rj.html

cheers

jonathon
User avatar
jonathon
 
Posts: 1282
Joined: Fri Jan 27, 2017 10:04 am

by ConalMonaghan » Tue Mar 03, 2020 1:29 am

Hi Jonathon and Devecct,
I think this would be a great addition to JAMOVI. Presently, we have to do all data cleaning and screening in R, then export into JAMOVI. While it is possible to do tests just as Mahalanobis (multi-variate outliers) using the work around in Rj and the psych package, there is no package in the JAMOVI database that can test for missing at random (MCAR), or engage in imputation or regression substitution.

Potentially having a module that tested the nature of missing data (MCAR) and imputation as needed would work well, alternatively, if you could introduce another package such as MissMech to the JAMOVI r library then the Rj solution would work.
ConalMonaghan
 
Posts: 1
Joined: Mon Mar 02, 2020 12:54 am

by jonathon » Tue Mar 03, 2020 1:33 am

what if you use 'system R', from Rj, rather than the R bundled with jamovi?

https://blog.jamovi.org/2018/07/30/rj.html

jonathon
User avatar
jonathon
 
Posts: 1282
Joined: Fri Jan 27, 2017 10:04 am


Return to General