Lok,
As Fredrik mentioned, the best way to -merge- the two dataset is by taking
the individual one as the "master" and the household as the "using" one.
This procedure will ensure that all the information will be in a single
dataset. From my point of view you should use this dataset at the individual
level even if your unit of analysis is the household.
If you prefer to work only with household data you have to -collapse- the
individual dataset into means, sums, etc. before merging it with the
household one. In this last case it wont matter which dataset you take as
the "master or the "using" one.
Rafa
----- Original Message -----
From: "Fredrik Wallenberg" <[email protected]>
To: <[email protected]>
Sent: Sunday, May 23, 2004 6:47 AM
Subject: Re: st: Questions on Merging Two Data sets(Individual and Household
data)
> On Sat, 22 May 2004 02:05:53 -0700, Lok Wong <[email protected]>
wrote:
> >
> > Hi there,
> >
> > I have two data sets: one individual, another household. The
observations in
> > household data set are less than those in individual one
> > because household consists of more than one individual. Now, I want to
merge
> > this two data sets. What should I do? In the invidiual data,
> > there are two variables "household id" and "number in the household"
which
> > could be used to identify invididuals uniquely. In the household
> > data, there is one variales "household id".
> >
>
> The first question is whether you want to make your analysis at the
> individual level or at the household level. The first is straight
> forward. Sort (and save) the household level data and then use and
> sort the individual data. After that you use
>
> merge -hhid- using -hhdata-
>
> This will add the hhdata to each individual observations with
> individuals in the same household having the same household level
> variables. The issue here is one of weighting but someone with a
> better grasp of statistics/econometrics can probably provide you with
> an appropriate correction.
>
> In the other case you would want to use some aggregate characteristics
> for the members of the household in household level data. Depending on
> the characteristics you have this may end up being an average, a
> count, a total or, in some cases, the characteristics associated with
> the head of the household (ignorining the other members of the
> household). You can create these variables (the first three at least)
> using -compress, by hhid- and then merge on -hhid-.
>
> I know this response isn't very deep, but the Stata implementation is
> straight forward once you know what you want to accomplish.
>
> Fredrik
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
