Statalist


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

st: R: Loss of observations with lagged variables in survival analysis


From   "Carlo Lazzaro" <carlo.lazzaro@tin.it>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: R: Loss of observations with lagged variables in survival analysis
Date   Tue, 8 Sep 2009 10:23:18 +0200

Dear Luis,
assuming that the last observation only is  missing for each individual (if
this is the case, your analysis would be affected by a limited bias), can't
you simply right-censor the survival analysis at wave #7 (provided that all
the individuals have the same probability of being right-censored)? 

Should the previous hint be unfeasible, I would refer you to the following
article, hopefully of some help:

Pfeifer PE, Bang H (2005) Non-parametric estimation of mean customer
lifetime value. Journal of Interactive Marketing. Doi:10.1002/dir.20049

Kind Regards,
Carlo

-----Messaggio originale-----
Da: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] Per conto di Luis Ortiz
Inviato: lunedì 7 settembre 2009 18.43
A: statalist@hsphsun2.harvard.edu
Oggetto: st: Loss of observations with lagged variables in survival analysis

I am using information on wages for an event-history (survival) analysis.
This information needs to be lagged forward. As a result, the latest
observation of this variable for each individual is missing. Here I post an
example for a given individual of my sample: 

pid	wave	wage	      wage_lag
20102	1	1577.2	1712.125
20102	2	1712.125	2242.275
20102	3	2242.275	1934.929
20102	4	1934.929	2298.635
20102	5	2298.635	2493.237
20102	6	2493.237	2667.933
20102	7	2667.933	3341.879
20102	8	3341.879	.

This has an effect on my analysis. Introducing this lagged variable in the
analysis automatically means losing observations, since the lagged variable
does not exist for the latest observation, as you may see above. This is
rather unfortunate, since some of these periods are the latest ones in my
spells of analysis. 

Is there any way of coming over this problem you come up with?

Many thanks for your attention

Luis 

-.-.-.-.-.-.-.-
Luis Ortiz
Profesor Agregado
Departament de Ciencies Polítiques i Socials
Universitat Pompeu Fabra
Ramon Trias Fargas, 25-27
08005 Barcelona
 
Phone: +34-93-5422368
Fax: +34-93-5422372
http://www.upf.edu/dcpis/
http://sociodemo.upf.edu/
 
 



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index