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

st: Multi-Level nlogit regression

From   Brendon Gerding <>
Subject   st: Multi-Level nlogit regression
Date   Wed, 15 Jul 2009 14:49:29 +0200

Hi, I am fairly new to Stata and am busy working on my first project
which is my honors dissertation. One part of the project is to
estimate whether an individual that died (ie an observation in the
mortality file of certain census data) was employed or unemployed, and
if they were employed, whether this employment was in the formal or
informal sector. In South Africa (I'm not sure about in other parts of
the world) employment status is not recorded with the deceased's
details, and it is therefore necessary to use regression to predict
employment status. I am using the person file (those individuals who
are still alive) in order to estimate regression coefficients to use
in predicting employment statuses of those observations in the
mortality file.

Since I am able to break down my response variable into a nested form,
I figured it would be best to use the nested logistic regression (we
were not taught this model, but from what I have read it seems to be
the most appropriate, and since other regressions have not worked, it
is my last hope).

The logical break-down of employment status would be as follows:

Population: -> Not Economically Active
                 -> Economically Active  -> Unemployed
                                                    -> Employed ->
Employed in formal sector

-> Employed in informal sector

How would I go about setting up this 3-level nlogit regression? And
then how would I use the postestimation results to predict employment
status of those individuals who have died?

Any help would be greatly appreciated.

Many Thanks,
Brendon Gerding
University of Cape Town Student
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index