[R-sig-ME] Multi-level Rasch Model Per Douglas Bates' paper

[Simon Harmel]

Dear Phillip,

The post you saw was about the "model" itself not the question I have asked
here:)

Dear Ben,
Thank you:) as you know my data is huge and won't fit any plot. I simply
want to make sure that "*item easiness estimates*" and "*person ability
estimates*" are correctly obtained in the following way for each model
given Prof. Bates' paper:


r11 <- ranef(m1, condVar = TRUE) # for 'm1' model in my original question
in this thread
r22 <- ranef(m2, condVar = TRUE)  # for 'm2' model in my original question
in this thread

### Person abilities and item easinesses for 'm1'
person_abilities11 <- r11$cond$person_id$`(Intercept)`
item_easiness11 <- r11$cond$item_id$`(Intercept)`

### Person abilities and item easinesses for 'm2'
person_abilities22 <- r22$cond$person_id$`(Intercept)`
item_easiness22 <- r22$cond$item_id$`(Intercept)`


On Wed, May 13, 2020 at 10:54 AM Ben Bolker <bbolker using gmail.com> wrote:

>     Yes, this is cross-posted, and I was planning on getting around to
> showing how to do it in the Cross-Validated post. glmmTMB doesn't have a
> built-in dotplot method, but you can cheat pretty easily because the
> individual components of a ranef() extracted from a glmmTMB fit ($cond
> for conditional model, $zi for zero-inflation model if any) have the
> same structure as ranef() from lme4, so you can steal the plotting method:
>
> library(glmmTMB)
> example(glmmTMB)
> library(lme4)
> r <- ranef(m1)
> library(lattice)
> lme4:::dotplot.ranef.mer(r$cond)
>
>    Note that it's not as easy to get dotplots with whiskers from coef()
> because of some long-standing (and deep) issues with computing standard
> deviations for the sum of a fixed and a random effect ...
>
> On 5/13/20 11:28 AM, Phillip Alday wrote:
> > I think I saw this go past on CrossValidated -- you should mention any
> > crossposting. :)
> >
> > In general, it would be nice to know what the structure of your data
> > are. Is "gender" a property of your participants, items, or something
> > else? What about item_type?
> >
> > In lme4, you can extract the item-level predictions with coef(m) (which
> > is the same as ranef(m) + fixef(m)).  You can even get a plot of these
> with:
> >
> > library(lattice)
> >
> > dotplot(ranef(m, condVar=TRUE))
> >
> > The zero-point is the grand mean (i.e. the corresponding fixed effect).
> >
> > I don't know if this is the same as in glmmTMB.
> >
> > Best,
> >
> > Phillip
> >
> > On 13/5/20 5:07 pm, Simon Harmel wrote:
> >> Hi All!
> >> I'm following this paper <
> https://www.jstatsoft.org/article/view/v020i02> (
> >> https://www.jstatsoft.org/article/view/v020i02) by Prof. Bates where
> after
> >> fitting the model (*pp. 14-15*), they obtain what they call *item
> >> easiness* *"from
> >> the estimates of the fixed effects and the conditional modes of the
> random
> >> effects."*
> >>
> >> In short, I wonder how to obtain item easiness estimates for each of my
> >> models (m1 & m2) below? *Thank you, Simon*
> >>
> >> library(glmmTMB)
> >> dat <- read.csv('https://raw.githubusercontent.com/ilzl/i/master/d.csv
> ')
> >>
> >> form11 <- y ~ item_type + (1 | item_id) + (1 | person_id)
> >>
> >> form22 <- y ~ item_type + gender + (1 | item_id) + (1 | person_id)
> >>
> >>
> >> m1 <-  glmmTMB(form11, data = subset(dat, person_id <= 40),
> >>                         family = beta_family())
> >>
> >> m2 <-  glmmTMB(form22, data = subset(dat, person_id <= 40),
> >>              family = beta_family())
> >>
> >>      [[alternative HTML version deleted]]
> >>
> >> _______________________________________________
> >> R-sig-mixed-models using r-project.org mailing list
> >> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
> > _______________________________________________
> > R-sig-mixed-models using r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
> _______________________________________________
> R-sig-mixed-models using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>

	[[alternative HTML version deleted]]