Page 1 of 1

EFA based on covariance/ unstandardized results

Posted: Wed Aug 17, 2022 11:13 pm
by Abdullah
Hi,

Jamovi runs the EFA nicely. However, in some cases, we want the covariance to read the unstandardized estimates rather than the correlation/standardized parameters. JASP does that already and we would love to see it in Jamovi.

Cheers

Re: EFA based on covariance/ unstandardized results

Posted: Tue Aug 23, 2022 8:56 am
by Ravi
Hi Abdullah,

Could you share the use case with us? Adding an options to do this is technically very easy, but we're always focussed on keeping an analysis as simple as possible so that beginner users are not overwhelmed by options that almost nobody uses. But of course, if this is an options that is very important to a lot of people we can be convinced to include it. In the end, it's a tight balance between being as easy to use as possible while still being as complete as possible.

Cheers,
Ravi

Re: EFA based on covariance/ unstandardized results

Posted: Tue Aug 23, 2022 2:17 pm
by Abdullah
Thank you for the reply, Ravi.

One can use these scores for different purposes. For me, at least, I need these scores as starting values on a subsequent analysis in SEM (through SEMLj Modul). We need the unstandardized results from the EFA to be used there. Also, they can be useful if the EFA includes measures with different lengths (e.g., a factor out of 5 while another out of 20). In the end, as long as they can be implemented easily, they will be useful for some users and at least, not harmful, I guess

Also, I wonder if more model fit indices can be incorporated in the EFA option "model fit measures" such as CFI and SRMR (from the CFA), they are certainly useful.
Many many thanks for your effort, Ravi

Re: EFA based on covariance/ unstandardized results

Posted: Sun Sep 11, 2022 6:00 pm
by Abdullah
Up