st: rasch models

["Henrique Neder" <hdneder@ufu.br>]

Dear members of the Stata mailing list

I'm working with estimates of the Rasch model for studies on
multidimensional poverty based on samples with probabilistic weights. My
intention is to test the existence of dimensions of poverty based on sets of
binary variables related to shortcomings of households , eg the household
does not have (or have) at least an occupied person with a income higher
than the minimum wage, or the household does not has (or has) a particular
appliance.
For this, I am using a set of commands in Stata to estimate these binary
Rasch models. However, the commands bachrasch and raschtest  do not allow
the use of weights to estimate the models . There is a possibility this
problem be solved? If not using the weights in the probabilistic estimation
would unduly estimating the dimensions of poverty, from the values of the
latent trait for each household?  Another issue is that I want to estimate
the model for each set of specific areas and thereafter obtaining the
estimates of latent traits for each of these regions separately. This
procedure is correct? How do I interpret the property (or result) of this
model called specific objectivity? I am grateful in advance for a possible
help regarding these issues.


Best regards,

Henrique Neder


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/