Timeline for How to fit Graded Response Model with lme4::glmer
Current License: CC BY-SA 3.0
15 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 1, 2015 at 12:58 | answer | added | KH Kim | timeline score: 1 | |
| Oct 12, 2014 at 18:41 | history | edited | KH Kim | CC BY-SA 3.0 |
added 1743 characters in body
|
| Sep 5, 2014 at 4:43 | history | edited | KH Kim | CC BY-SA 3.0 |
added 133 characters in body
|
| Sep 5, 2014 at 4:37 | history | edited | KH Kim | CC BY-SA 3.0 |
added 133 characters in body
|
| Sep 5, 2014 at 4:29 | history | edited | KH Kim | CC BY-SA 3.0 |
added 655 characters in body
|
| Sep 5, 2014 at 3:53 | comment | added | KH Kim | @Jake Westfall, Rijmen et al.(2003) says, "The mixed logistic regression model can handle polytomous responses by ..." and I thought mixed logistic regression model can be covered by GLMM... | |
| Sep 5, 2014 at 1:23 | comment | added | KH Kim | @Jake Westfall, I guess I am wrong about the way of implementing my data. For each category, the code should be 0 and 1 and the only category the subject selected should be 1. and the logit function should be different, that is cumulative logit function. So I guess I have to redefine link function...??? | |
| Sep 5, 2014 at 1:15 | comment | added | KH Kim | @Jake Westfall, As far as I figure, he is alluding that Polytomous Data can be modeled by GLMM(Generalized Linear Mixed Model, I am not referring to any particular package here). See page 191 of Rijmen et al.(2003) $L_{nij}=x_{nij} beta + z'_{nij} theta _n$ So everything's fine here. Link function and linear predictors | |
| Sep 5, 2014 at 1:14 | comment | added | KH Kim | @Momo, I guess you're right. Sorry for not mentioning. I thought of GRM with constrained slope parameter. | |
| Sep 4, 2014 at 16:56 | comment | added | Jake Westfall |
1st comment: As far as I know, Rijmen et al. do not say that one can fit a GRM using lme4::glmer(), and indeed I don't see how one could--please provide a page reference in Rijmen et al. if you disagree. 2nd comment: I think you can fit GRM in ordinal::clmm(), although I haven't personally looked into it closely. 3rd comment: I don't know what your actual question is. Please edit your question to very clearly state what it is you want to know or are confused about.
|
|
| S Sep 4, 2014 at 16:45 | history | suggested | Steve | CC BY-SA 3.0 |
improved formatting
|
| Sep 4, 2014 at 16:36 | review | Suggested edits | |||
| S Sep 4, 2014 at 16:45 | |||||
| Sep 4, 2014 at 16:30 | comment | added | Momo | I'm pretty sure you can't fit GRM in glmer, as the gl stands for generalized linear and the GRM is a generalized nonlinear model. This means the structural part of the model is not a linear predictor but is of the form $a*(b+c)$ with $a,b,c$ unknown. Linear would be something like $b+c$ (e.g., a Rasch model). I therefore agree with de Boeck. | |
| Sep 4, 2014 at 15:45 | history | edited | KH Kim | CC BY-SA 3.0 |
added 363 characters in body
|
| Sep 4, 2014 at 15:24 | history | asked | KH Kim | CC BY-SA 3.0 |