[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: new package seqlogit available on ssc

From   Maarten buis <>
Subject   st: new package seqlogit available on ssc
Date   Thu, 16 Aug 2007 12:25:35 +0100 (BST)

Thanks to Kit Baum a new package called -seqlogit- is now available on
ssc. This package fits a sequential logit model. Traditionally, the aim
of a sequential logit model is to study a process that can be described
as a series of choices between a small number of options. However,
these choices eventually lead to an end result. The aim of this package
is to not only study the effects of explanatory variables on these
choices (study the process), but also study the effects on the end

For example, in education parents make a series of choices at what type
of schools to send their children.  Typically, the aim of the model is
explore the effects of explanatory variables on the probabilities of
continuing from one level to the next, i.e. study the process through
which education is attained. However, in the end this process lead to
an outcome: highest achieved level of education. Both the process and
the end result are of substantive interest and are related to one
another, but they are not the same.

This package implements a method by Buis (2007) to study simultaneously
the effect of explanatory variables on the process (the probability of
passing from one level to the next) and the final outcome (highest
achieved level). The effect of the explanatory variable on the end
result can be shown to be a weighted sum of the log odds of passing the
transitions. The weights are the product of the following three

o the proportion of people at risk of passing the transition, so a
  transition receives more weight if more people are at risk of passing
o the variance of the dummy indicating whether the transition was
  passed or not, so a transition receives more weight if close to 50%
  pass, and less weight if virtually everybody passes or fails that
o the expected difference in outcome between those that pass and those
  that fail the transition, so a transition receives more weight if
  people gain more from passing it.
-seqlogit- will estimate a sequential logit and allows one to use
-predict- to predict the weights, and its components. Furthermore, it
contains the -seqlogitdecomp- command, which graphically displays the
decomposition of the effect of the explanatory variable on the end
result into the effects of the explanatory variable on the individual
transitions and their weights.  

This package also contains three ancillary files, that describe an
example. The files are a do-file that implements the example, the
dataset used by the do-file, and a pdf-file which shows the output and
explains some of the tricks used in the example. To install the package
and the ancillary files type -ssc install seqlogit, all-

I will give a presentation on this package on Friday September 7 at the
2007 Nordic and Baltic Stata Users Group meeting. 

What I call here a sequential logit model is also known under a number
of different names: sequential response model (maddala 1983),
continuation ratio logit (Agresti 2002), model for nested dichotomies
(fox 1997), and the Mare model (shavit and blossfeld93) (after (Mare


Agresti, Alan 2002 .  Categorical Data Analysis, 2nd edition.  Hoboken,
NJ: Wiley-Interscience.

Buis, Maarten L. 2007 ``Not all transitions are equal: The relationship
between inequality of educational opportunities and inequality of
educational outcomes''

Fox, John 1997 Applied Regression Analysis, Linear Models, and Related
Methods.  Thousand Oaks: Sage.

Maddala, G.S. 1983 Limited Dependent and Qualitative Variables in
Econometrics Cambridge: Cambridge University Press.

Mare, Robert D. 1981 ``Change and Stability in educational
stratification'' American Sociological Review, 46(1), p.p. 72-87.

Shavit, Yossi and Hans-Peter Blossfeld 1993 Persistent Inequality:
Changing Educational Attainment in Thirteen Countries Boulder: Westview

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Yahoo! Answers - Got a question? Someone out there knows the answer. Try it
*   For searches and help try:

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