[R] randomLCA

Gabriele Accetta gabriele.accetta at gmail.com
Tue Aug 28 13:29:56 CEST 2012


Dear Ingmar, thank you for your email.

This means that a non-random latent class model for a 2x2x2 table (3
diagnostic tests) and 2 latent classes produces a saturated model. The
predicted frequencies equal the observed frequencies.

Is there any reason to introduce  random effects in a saturated model?

thanks
gabriele

2012/8/28 Ingmar Visser <i.visser at uva.nl>:
> use str(dentistry.lca2) to see all values of the output; among them a value
> np for number of parameters, in this case 5*2 for the 5 binary items of 2
> classes + 1 for the class proportions, total 11.
> hth, Ingmar
>
> On Mon, Aug 27, 2012 at 6:05 PM, Gabriele Accetta
> <gabriele.accetta at gmail.com> wrote:
>>
>> Can anybody, please, explain me how many parameter are estimated using
>> randomLCA?
>>
>> For examples,  model  "dentistry.lca2random"   estimate 1 scale (or
>> variance, b_j)  parameter  and 2 position parameters (a_cj)? Doesn't
>> it?
>> Do I need at least 4 diagnostic tests for such a model?
>>
>> What happens if I specify options blocksize and byclass? How many
>> diagnostic tests (or rater) I need?
>>
>>
>> Extract from see "randomLCA examples", by Ken Beath.
>>
>> > dentistry.lca2random <- randomLCA(dentistry[,
>> + 1:5], freq = dentistry$freq, initmodel = dentistry.lca2,
>> + nclass = 2, random = TRUE, quadpoints = 31,
>> + probit = TRUE)
>>
>>
>> Thank you.
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide
>> http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>
>



-- 
Gabriele Accetta

"E' evidente che non si ha vera democrazia là dove l'accesso
all'istruzione non è garantito in misura pari a tutti i cittadini:
perché, importando necessariamente la diversa cultura una diversa
possibilità di partecipazione alla vita politica, il privilegio
dell'istruzione si risolve necessariamente in privilegio politico."
Piero Calamandrei




More information about the R-help mailing list