--- Ziad El-Khatib <[email protected]> wrote:
> I have similar situation to yours (for unknown blood results) and
> code them 999999999999.
> I so as Sergiy suggested or add 'if var!=999999999999' (which can be
> annoying sometimes).
You can simplify that (and thus avoid typos) in a number of ways:
1) Keep the raw data asis, but create a .do-file called dataprep.do
that will contain a set of commands like this:
*----------- dataprep.do -----------------
use rawdata.dta, clear
recode var = . if var == 999999999999
assert var < 999999999999 | var == .
other manipulations of the data
*----------- end dataprep.do ------------
Notice that dataprep.do does not save the "cleaned" data. This is also
a good place to add some checks like the -assert- command. I would also
liberaly add comments to document what you are doing to your data and
why. Than you write another .do file for your analysis, call it
analysis.do, and it would read like this:
*----------- analsysis.do -------------
do dataprep.do
sum var
reg y x var
some more wonderful analyses
*------------ end analysis.do ---------
2. You can use -mark- as described in (Jann 2007):
mark touse if var == 999999999999
sum var if touse
reg y x var if touse
Hope this helps,
Maarten
Ben Jann (2007) Stata tip 44: Get a handle on your sample. The Stata
Journal, 7(2):266--267.
http://www.stata-journal.com/article.html?article=dm0030
-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands
visiting address:
Buitenveldertselaan 3 (Metropolitan), room N515
+31 20 5986715
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
*
* 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/
