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

st: Estimating Period Effects in Repeated-Measures Analysis

From   Kristine Witkowski <>
Subject   st: Estimating Period Effects in Repeated-Measures Analysis
Date   Mon, 10 Feb 2003 14:48:06 -0500


Using 3 waves of repeated-measures from a panel sample,
I want to estimate whether the likelihood of my outcome
(being depressed) depends on which wave this information is
collected, controlling for a variety of time-varying covariates.

I generate three indicators where time1=1 for Wave-1 data,
time2=1 for Wave-2 data, and time3=1 for Wave-3 data
(with corresponding zeros assigned for each indicator when
data is not from the wave of interest).

I then use logistic regression for repeated measures (stacked data file)
where the sytnax is

  logistic majdep time2 time3 indvar, cluster(id)

My period indicators (time1,time2,time3) covary with one another
because the nature of panel data and therefore the model drops all
but one of my period indicators.  Those in the panel at Wave-3
(where time3=1) must also be in the panel at Waves-1,2 (time1,2=1).

I should I model things differenty?

Thank you for your assistance in advance.


Kristine Witkowski, PhD			
The University of Michigan
School of Social Work
1080 South University, B660-22 SSWB
Ann Arbor, MI  48109-1106
Phone: 734-615-2101
Fax:  734-615-7742

*   For searches and help try:

© Copyright 1996–2019 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index