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: Recode and Label problem


From   daniel klein <klein.daniel.81@googlemail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Recode and Label problem
Date   Wed, 17 Oct 2012 09:02:03 +0200

Malay,

I have a bundle of programs -labutil2- (SSC) that deal with value
labels. There are several solutions to your problem with those
programs (including a recently programmed,  but not yet released), but
I think in your case you might be just as fast with official Stata
commands.

First off, note that in your case

recode <foo> (2=0)

will change the original variable. That is, in most cases not a good
idea. I will, however ignore that for now. Note that this does _not_
change any labels whatsoever. I just wanted to point out a faster
solution

replace <foo> = <foo> - 1 if !mi(<foo>)

will give the same binary indicator, but -replace- (or -generate-, if
you want a new variable) is much faster than -recode- (might not be a
problem in a simple do-file, though).

I think a convenient way to solve your problem, is to use one value
label, say <yesno>, for all your binary indicator variables. Given you
already have all your variables changed to indicators, you type

la de yesno 1 yes 0 no
la val <firstindicator>-<lastindicator> yesno

to do this. If you want to create new variables, using -recode- would
be convenient (also Nick Cox might disagree here)

recode <firstvartochange>-<lastvartochange> (1 = 1 "yes")(2 = 0 "no")
,pre(b_) l(yesno)

Omitting option -label()- will create a value label b_<varname> for
each of the new indicator variables.

Best
Daniel

-- 
Dear All,
Thank you for all the supports given to me by some of you in my
earlier requests. Sincerely appreciated.

I am having one small but nagging problem. I know that I am just one
small command away of what I want - but it had alredy consumed quite a
bit time.

Problem:
Good number of variables in my data were coded as 1( Yes) and 2 (No).
(Problematic for logistic analysis)

So I used the command: recode var39 2=0   (This command changes all my
values (2) in the variable39 to 0. But while executing the command, it
changes my label too and dsiplay 0. So again I use :

label define var39 0 "No" 1 "Yes"

It does not change the label in the data from zero to No. In all my
analysis it shows Yes as Yes but No to 2. What I want is to code it as
zero but label it as No.
(var39 is my variable having two options: Yes(1) or No (2))

I am not sure whether I have articulated the problem well.

Thank you for your time.
With regards
Malay
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index