Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

st: -xtmixed- and multilevel data [Was: Grouping income variables- RECODE COMMAND]


From   <[email protected]>
To   <[email protected]>
Subject   st: -xtmixed- and multilevel data [Was: Grouping income variables- RECODE COMMAND]
Date   Wed, 5 Feb 2014 10:07:44 +0000

In addition to having issues with how to deal with grouped income data with missing values, I note that Antonio Rodriguez Andres <[email protected]> is using European Social Survey data -- pooling data for 23 countries, with a large number of respondents per country, and then fitting a mixed model.

I suggest that Antonio proceed with caution with this modelling approach. The number of countries is "small" and so country effects (coefficients on country fixed effects or random effect variances) may not be reliably estimated.

For a review of the issues, see:

'Regression analysis of cross-national differences using multi-level data: a cautionary tale', Working Paper 2013-14. Colchester: Institute for Social and Economic Research, University of Essex. https://www.iser.essex.ac.uk/publications/working-papers/iser/2013-14 

Related: see also "A Monte Carlo analysis of multilevel binary logit model estimator performance" presented at the 2013 UKSUG: http://www.stata.com/meeting/uk13/abstracts/ 


Stephen
------------------
Stephen P. Jenkins <[email protected]>

Date: Tue, 4 Feb 2014 14:48:54 +0200
From: "Antonio Rodriguez Andres" <[email protected]>
Subject: RE: st: Grouping income variables- RECODE COMMAND

Dear Maarten

Thank you very much for your feedback. What I did is the following

http://www3.nd.edu/~rwilliam/stats2/l12.pdf

*Create income midpoints

recode hinctnt (1=900) (2=2700) (3=4800) (4=9000) (5=15000) (6=21000) (7=27000) (8= 33000) (9=48000) (10=75000) (11=105000) (12= 175200) , gen(hincome)
replace hincome=. if hinctnt==77 | hinctnt==88 |  hinctnt==99
gen lhincome=log(hincome)

**dummy indicator for missing income values

gen xhincome=hincome
replace xhincome= 29304.99 if missing(hincome)
gen md=0
replace md=1 if xhincome! =hincome

xtmixed dprt age age2 gender married separated divorced widowed eduyrs ichldhm md lhincome ihealth iuemp5yr iuemp12m rgdp06[pw=dweight] if md==0 || cntry: gender , mle


But I still got the same message, the md indicator variable is dropped. How can Ä° estimate the model controlling for missing values in income?

Antonio

Please access the attached hyperlink for an important electronic communications disclaimer: http://lse.ac.uk/emailDisclaimer

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


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index