Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: Latent Variable and IV


From   "Mafalda Sampaio" <mafaldasampaio@fe.unl.pt>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: Latent Variable and IV
Date   Mon, 8 Feb 2010 20:19:49 -0000

Hi,
Thank you for your quick answer. I'm studying how people choose their networks, and one of the variables is the "perceived information". People interact with a potential friend and will or will not choose  to establish a link depending on what information he can extract from it. This perceived level of information is a latent variable, as it is not observable but depends on observables. I'm thinking of recovering with the use of some indicator variable say, "whether or not the potential friend has access to priviledge information" or something of the sort (data that I have)...
As for the IV, I have another explanatory variable that is endogenous. If I  had not the latent variable I would use 2SLS. However, having these two problems with my covariates, I'm not sure what to use.

I hope I was clearer...

Thank you!

Mafalda

-----Mensagem original-----
De: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] Em nome de Maarten buis
Enviada: segunda-feira, 8 de Fevereiro de 2010 20:13
Para: stata list
Assunto: RE: st: Latent Variable and IV

--- Mafalda Sampaio wrote:
> In my research I have a model that has both a latent
> explanatory variable and a endogenous variable. So, in
> estimating I have to tackle both. 
>
> I did not seem to find a fancier way to estimate this so I
> thouhgt of doing a "by hand" first stage, estimating the 
> endogenous variable through its instrument, and then
> applying the estimated variance on the model with the
> latent variable. However, this would lead to wrong SE. 
>
> Does anyone has an idea? Does stata deals simultaneously
> with IV and Latent? 

A latent variable is just a variable hat has not been 
directly observed. So, you need some other strategy to 
recover it from observed data. There are many strategies,
which are each appropriate to a different set of problems.
So just saying that you have a latent variable does not
give us enough information to answer your question: So,
what kind of latent variable do you have in mind, what
kind of strategy do you want to use to recover it, what
kind of IV estimation do you want to perform?

-- Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
No virus found in this incoming message.
Checked by AVG - www.avg.com 
Version: 8.5.435 / Virus Database: 271.1.1/2673 - Release Date: 02/08/10 07:35:00


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



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index