Re: st: Data conversion from plot level to household level

[Phil Schumm <pschumm@uchicago.edu>]

On Aug 30, 2007, at 4:57 AM, Sridhar Thapa wrote:
> I want to construct data  at household level from plot level and  
> individual level.  I was thinking to do it manually but it is  
> highly time consuming because of large data set.

Not only time consuming, but error-prone.  Don't do that.


> I did not see any command to manage it at household level. Could  
> anyone give me some ideas how to manage it? I would be indebted  
> with you. Thanking you in advance.
> Data set like this(HH is the identifier for household level)
> HH   Plotno      Crop	labour male  labour female
> 01	1	Paddy	30mandays	35
> 01	2	Maize	20	        22
> 01	3	wheat	12	        16
> 02	1	Paddy	16	        28
> 02	2	maize	15	        18
> I intend to convert these plot level data as an aggregate household  
> level data set

Two basic strategies here.  The first is to create your aggregate  
variables -- all constant within household -- and then drop all but  
one observation per household.  There are lots of ways to create  
aggregate variables depending on your requirements, but -egen- with  
the -by()- option offers a lot in this regard, so that would be a  
good place to start.  One you've created your aggregate variables,  
you can use

bys hh: keep if _n==1

to drop all but one observation per household.  A safer way would be  
first to drop all of the variables except (1) household id, and (2)  
your household-level aggregate variables, and then use

duplicates drop
isid hh

which, if it works without error, will verify that you created your  
household aggregates so that they are indeed constant within household.

A second strategy is to use -collapse-, which can do all this in one  
step.  The downside then is that you're limited to the aggregate  
statistics that -collapse- provides.  However, if those suit your  
needs, then this is the more efficient strategy.


-- Phil

*
*   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/