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

Re:st: How to weight datasets?

From   "rafaelcla" <[email protected]>
To   "statalist" <[email protected]>
Subject   Re:st: How to weight datasets?
Date   Sat, 30 Aug 2008 21:46:37 -0300


In order to obtain weighted estimations � and consequently fit your data on a specific population � you�ll need to �place the weight� on estimation command NOT on the data. �To Weight� the data (like, by replicate a single case X times according to its weight) does not work in many, many situations�and will possibly drive you to wrong results�so, don�t do this unless you are totally sure about it.

As I�m not exactly sure about the sample design (and the kind of weight) you have�I�ll give you an example as if you were using �probability weights�. After loading your dataset, to obtain a weighted mean of variable �X�, using the weight variable �Y� (that should indicate the probability weight or just PW), all you need to do is type:

mean X [pw=Y] 

In fact, as you can note, the �weighting procedure� (described by the extension �[pw=Y]� on the command above) works as an extension of the command�not a command itself (if you type just "mean X"stata will give you the unweighted mean). 

Not all kinds of weight (PW, AW, IW�) work with all commands�but you can figure this out on the help file of the commands you want to proceed with. 

Rafael Claro

---------- In�cio da mensagem original -----------

      De: [email protected]
    Para: [email protected]
    Data: Sat, 30 Aug 2008 15:24:35 +0200
 Assunto: st: How to weight datasets?

> I have four separate datasets that I would like to weight in order to fit them on a specific population. I know that Stata has commands like pweight or aweight, but how exactly do I use them to create a weighted dataset?
> Maybe someone has a do-file to show me or simply write me the needed commands. The Stata- help is no good for me, since I`m not able to use it correctly. 
> Thanks a lot
> -- 
> GMX startet Hier findest Du Leute mit Deinen Interessen!
> Jetzt dabei sein:
> *
> *   For searches and help try:
> *
> *
> *

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index