save factor scores, regressions residuals, etc. ?

Discuss the jamovi platform, possible improvements, etc.

by jonathon » Fri Dec 14, 2018 9:57 pm

*chuckles*, that's true!
User avatar
jonathon
 
Posts: 841
Joined: Fri Jan 27, 2017 10:04 am

by Folax » Mon Dec 17, 2018 1:51 pm

South winter 2019 would be fine ! ;-)
Folax
 
Posts: 8
Joined: Tue Sep 18, 2018 2:29 pm

by jonathon » Tue Dec 18, 2018 12:05 am

haha, we'll see what we can do :)
User avatar
jonathon
 
Posts: 841
Joined: Fri Jan 27, 2017 10:04 am

by Wake » Mon May 06, 2019 7:36 am

Please could we have an update? Is this still likely for September 2019?

Cheers,

Wake
Wake
 
Posts: 15
Joined: Tue Jun 26, 2018 8:31 am

by jonathon » Mon May 06, 2019 7:37 am

i'll need to discuss it with the team. let me get back to you.

jonathon
User avatar
jonathon
 
Posts: 841
Joined: Fri Jan 27, 2017 10:04 am

by jonathon » Mon May 06, 2019 7:42 am

i think it may be pushing it. is there a possibility your uni could perhaps sponsor it?

jonathon
User avatar
jonathon
 
Posts: 841
Joined: Fri Jan 27, 2017 10:04 am

by Wake » Fri May 10, 2019 10:46 pm

It might be possible to Oxford Brookes University, or the Association of Technical Staff in Psychology to sponsor a component or two, but probably not until we've already adopted jamovi for undergraduates.

However, I've just finished my second step-by-step guide on Principal Components Analysis which includes a method for saving the component scores using the Rj Editor.

Best wishes,

Wakefield
Wake
 
Posts: 15
Joined: Tue Jun 26, 2018 8:31 am

by Wake » Sun May 12, 2019 10:07 pm

It looks like I'll have to rethink this one, for two reasons:

1) Whereas varimax rotation gives the same component loadings in jamovi and SPSS, oblimin seems to be very different.

2) The csv file being saved by jamovi seems to have lots of repeated scores -- it doesn't match SPSS, looks suspicious, and doesn't seem to change when switching between varimax and oblimin.
Wake
 
Posts: 15
Joined: Tue Jun 26, 2018 8:31 am

by jonathon » Mon May 13, 2019 12:12 am

1) Whereas varimax rotation gives the same component loadings in jamovi and SPSS, oblimin seems to be very different.


our experience is that you shouldn't assume that spss is correct.

2) The csv file being saved by jamovi seems to have lots of repeated scores -- it doesn't match SPSS, looks suspicious, and doesn't seem to change when switching between varimax and oblimin.


as in the .csv file of the scores computed by the R code in your slides?

jonathon
User avatar
jonathon
 
Posts: 841
Joined: Fri Jan 27, 2017 10:04 am

by Wake » Mon May 13, 2019 7:40 am

David Foxcroft spotted an error in my code. I had c(45:63) in the header code but c(45,63) in the actual PCA, so only two variables were included.

SPSS now matches jamovi for the varimax rotation so I'm not concerned about slight differences in the oblimin rotation.

I've updated my slides.

Cheers,

Wake
Wake
 
Posts: 15
Joined: Tue Jun 26, 2018 8:31 am

PreviousNext

Return to General