Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

Re: st: command stata

From   Richard Williams <>
Subject   Re: st: command stata
Date   Fri, 26 Aug 2011 09:02:40 -0500

Adding to Maarten's suggestions, the new -sem- (Structural Equation Modeling) module in Stata 12 makes it very easy to estimate MIMIC modules. See example 10 in the sem manual. I wish -sem- had been around 30 years ago back when I was struggling with the classic LISREL program.

To offer further insights on the original question, we really do need to know how the original variables are coded and what Donsaane means when she says she wants to "group them into one variable."

At 02:46 AM 8/26/2011, Maarten Buis wrote:
2011/8/26 Donsaane Dontsi Saatena :
> I am looking for stata's command in order to group many variables into one.
> Let's say we have: roads, water , electricity, and we want to group them
> into one variable: infrastructures.

--- Rich Williams proposed
> Does -egen- with the group function do what you want?

What Rich proposes is in essence to create one categorical variable
that has a unique value for each combination of values on each of the
three variables. I read this question differently. I interpret it that
you want to make a metric variable that indicates how good the
infrastructure is.

Here there is a subtle but important distinction between whether you
believe that the observed variables influence the latent variable or
the other way around. Consider an intelligence test; here we believe
that there is some latent intelligence that makes people answer
questions right or wrong, so the observed variables (whether or not
questions were answered correctly) are caused by the latent variable.
In your case it makes more sense to think of roads, water and
electricity as feeding into a pool of resources that we labelled
infrastructure, so now it is the observed variables that cause the
latent variable. In the former situation you would use factor analysis
and their kin, in the latter (your) situation you would use sheaf
coefficients, models with parametrically weighted covariates or MIMIC
models. These have been implemented in Stata in the -sheafcoef- and
-propcnsreg- packages, which can be obtained from SSC by typing in
Stata -ssc install sheafcoef- and ssc install propcnsreg-. You can
find out more about them here:

Hope this helps,

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
*   For searches and help try:

Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index