negative R2

[Filippo Santini]

Hi everyone,

I'm running a  model with some regressions between latent variable that go in both directions. I mean that the latent variable A is regressed on the latent variable B, and B is regressed on A.

I get negative R2 for some of the latent variables. Is it possible? Am I doing something wrong?

Thank you all

Filippo

[Terrence Jorgensen]

I'm running a  model with some regressions between latent variable that go in both directions. I mean that the latent variable A is regressed on the latent variable B, and B is regressed on A.

I get negative R2 for some of the latent variables. Is it possible? Am I doing something wrong?

This is a non-recursive model, which is prone to identification and estimation problems.  It is not a lavaan issue, so I would recommend asking for help on the more general SEMNET forum:

http://www2.gsu.edu/~mkteer/semnet.html

Terrence D. Jorgensen

Postdoctoral Researcher, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam

UvA web page: http://www.uva.nl/profile/t.d.jorgensen

[Filippo Santini]

Thank you Terrence!

[Edward Rigdon]

You may not be doing anything wrong, and your model may not have an identification or other problem. There is a problem at the core of the notion of R2 itself. The basic R2 formula calculate R2 as 1 - (error variance / total variance). This formula assumes a neat division between error variance and non-error variance. With nonrecursive models, this neat division does NOT hold, and R2 calculated using the standard formula can be just anything. There is no perfect fix, but Les Hayduk has popularized a modified R2 formula--a "blocked R2"--that produces better-looking results. It's not perfect--there doesn't seem to be a perfect solution, honestly.

https://link.springer.com/article/10.1007/s11135-005-1095-4

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
To post to this group, send email to lav...@googlegroups.com.
Visit this group at https://groups.google.com/group/lavaan.
For more options, visit https://groups.google.com/d/optout.