Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Tobit using svy commands

From   "Reza Daniels" <[email protected]>
To   <[email protected]>
Subject   st: RE: Tobit using svy commands
Date   Tue, 2 Mar 2004 14:50:51 -0800

Dear Adam,

Be sure when using -svyintreg- that you are aware of the differences between
a standard Tobit model (for example, in Greene WH, 2000: "Econometric
Analysis 4th Edition", Pg 911) and the interval regression model for which
-svyintreg- was written (see Stata manuals). In addition, note that when
incorporating complex survey design features into your model, it no longer
becomes possible to interpret the coefficients of the -svyintreg- results in
the same way as a Tobit model (see also Maddala, 1983, Pg 160), since your
likelihood function becomes a pseudo likelihood.

My former colleague and I have used -svyintreg- to estimate an earnings
function where the dependant variable (income) was grouped into intervals
(e.g. $1000-$2000), had actual (point) observations and zero (left censored)
observations. The strength of -svyintreg- is that all of these data types
can be accommodated, but in doing so the model changes to a generalisation
of the Tobit model and the equations differ in important ways.

Reza C. Daniels
Ph.D. Candidate, School of Economics
University of Cape Town, South Africa

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Adam Jayson
Sent: 02 March 2004 12:42
To: [email protected]
Subject: st: Tobit using svy commands

Dear Statalist:

I am trying to run a Tobit model regression using the svy commands in Stata
8, and I'm having trouble figuring out how to do this correctly.

My dependent variable is an expenditures variable; most of the values are
zero, but a third or so are in dollars greater than zero.  A Tobit model
seems like the right choice.  Given the design of the survey, I need to use
the svy commands.

As best as I can tell, I should be using the "svyintreg" command.  But that
command requires me to use two dependent variables, and I don't understand
what Stata's help file is telling me to do.  Is my real dependent variable
the first or the second of these two variables?  And what should the other
one be?

Thank you for your help.


*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index