Can blavan handle categorical data?

408 views
Skip to first unread message

Sarah Alsaadi

unread,
Mar 19, 2018, 9:55:02 AM3/19/18
to blavaan
Thanks in advance!

Best,

Sarah

Ed Merkle

unread,
Mar 19, 2018, 11:15:00 AM3/19/18
to Sarah Alsaadi, blavaan
Sarah,

blavaan cannot currently do this, but it is on the to-do list. Right now, your best bet is to export the JAGS or Stan code, then edit the code to handle ordinal data. But this admittedly requires some expertise.

Best,
Ed


On Mon, Mar 19, 2018 at 8:55 AM, Sarah Alsaadi <sarah....@gmail.com> wrote:
Thanks in advance!

Best,

Sarah

--
You received this message because you are subscribed to the Google Groups "blavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blavaan+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/blavaan/6fdf8831-443f-4ed2-9ff7-fbdbb007c274%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Message has been deleted

Terrence Jorgensen

unread,
Mar 20, 2018, 12:23:37 PM3/20/18
to blavaan
Right now, your best bet is to export the JAGS or Stan code, then edit the code to handle ordinal data. But this admittedly requires some expertise.

For guidance, you can find a description in chapter 5 this textbook, along with some example BUGS syntax (which is the same as JAGS syntax).


Terrence D. Jorgensen
Postdoctoral Researcher, Methods and Statistics
Research Institute for Child Development and Education, the University of Amsterdam

Mauricio Garnier-Villarreal

unread,
Mar 21, 2018, 7:41:38 PM3/21/18
to blavaan
Sarah

I had worked an example for the Graded Response Model (GRM) using JAGS. GRM is an IRT model for ordered categorical variables.

In this link I show my solution with JAGS for multidimensional GRM, as Ed said you could work from these code or blavaan exported code to adjust it for your case

Chris Spencer

unread,
May 3, 2018, 11:38:42 AM5/3/18
to blavaan
Hi All,

I have a follow-up question from this thread.  When you say that Blavaan does not currently support categorical variables, does this include exogenous covariates as well?

Thanks,
Chris


On Monday, March 19, 2018 at 11:15:00 AM UTC-4, Ed Merkle wrote:
Sarah,

blavaan cannot currently do this, but it is on the to-do list. Right now, your best bet is to export the JAGS or Stan code, then edit the code to handle ordinal data. But this admittedly requires some expertise.

Best,
Ed

On Mon, Mar 19, 2018 at 8:55 AM, Sarah Alsaadi <sarah....@gmail.com> wrote:
Thanks in advance!

Best,

Sarah

--
You received this message because you are subscribed to the Google Groups "blavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blavaan+u...@googlegroups.com.

Terrence Jorgensen

unread,
May 4, 2018, 7:25:23 AM5/4/18
to blavaan
When you say that Blavaan does not currently support categorical variables, does this include exogenous covariates as well?

No, there are no distributional assumptions about exogenous variables, so it is arbitrary whether they are continuous or discrete.  For proper interpretation, it is up to you to include dummy codes that make sense.
Reply all
Reply to author
Forward
0 new messages