Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: Use or don't use weights for combined cross-country observations in multi-level model


From   Clive Nicholas <clivelists@googlemail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Use or don't use weights for combined cross-country observations in multi-level model
Date   Sat, 25 Jun 2011 04:44:49 +0100

Melanie Kolbe wrote:

> after researching different opinions on the topic, I am still at a loss whether I should use weights or not for my model. I am interested whether being in the social group of interest (in comparison to everyone who is not)  affects their response to specific items from the survey I use. My observations are drawn from several EU countries as one country usually does not have enough respondents of the group of interest surveyed and I am also interested in the effects of country-level variables on the responses. So I am not comparing across countries, but just want to run a multi-level model in order to find out whether being in the group of interest affects the response option significantly or not. I use standard control variable such as gender, age, education, political alignment and so on in the regression.
> I have read that for regressions the use of weights can be messy and I am not sure that what I am doing requires weights at all. If anyone could give me a second or third opinion on that I would be grateful.

My only answer to your dilemma is this: just try it.

If your weights don't change your parameter estimates, standard errors
etc that much (and you're confident that your model tells a story
consistent with the theory or theories that you're testing), then the
chances are that you have little to worry about.

So, in short: (1) run the model; (2) ask questions afterwards; (3) see
if you can improve it.

-- 
Clive Nicholas

[Please DO NOT mail me personally here, but at
<clivenicholas@hotmail.com>. Please respond to contributions I make in
a list thread here. Thanks!]

"My colleagues in the social sciences talk a great deal about
methodology. I prefer to call it style." -- Freeman J. Dyson

*
*   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   |   Site index