R help - Feb 2011 - using the bootstrap method for latent class analysis

Hi

I was wondering whether anyone can provide any help or suggestions on the
boostrap method for latent class analysis please.

My main question is: can the bootstrap procedure sometimes be less precise
than the non-bootstrap procedure when carrying out latent class analysis?

I am asking this question because I carried out some analyses and when I *did
not* use the bootstrap procedure I found adequate fit for one of the
models but when I did use the bootstrap none of the models indicated
adequate fit to the data. I know that usually the bootstrap is used with
sparse data, however, I'm not sure whether I should use it or not in my
case.

I would greatly appreciate any feedback!!

Thanks!!


Debbie

	[[alternative HTML version deleted]]

>
> My main question is: can the bootstrap procedure sometimes be less 
> precise than the non-bootstrap procedure when carrying out latent class 
> analysis?
> I am asking this question because I carried out some analyses and when I 
> *did not* use the bootstrap procedure I found adequate fit for one of 
> the models but when I did use the bootstrap none of the models indicated
> adequate fit to the data.
It should not usually lead to this behaviour, though I am curious what 
model fit measure you are using in the bootstrap situation.  You need to 
tell us what packages you are using, and send us a small self-contained
example.

Cheers, David Duffy.

-- 
| David Duffy (MBBS PhD)                                         ,-_|\
| email: davidD at qimr.edu.au  ph: INT+61+7+3362-0217 fax: -0101  /     *
| Epidemiology Unit, Queensland Institute of Medical Research   \_,-._/
| 300 Herston Rd, Brisbane, Queensland 4029, Australia  GPG 4D0B994A v