I agree. That is, there would be need to be a way
to tell Stata about the mapping, as it would be
unfair to expect Stata to infer rules we didn't tell
it. And the way seems to be renaming, which is where
we were a while back.
Whether the variables mean much is a totally different issue.
Nick
[email protected]
Ada Ma
> Thanks Nick for the answer. I asked becoz I was thinking that there
> may be a way to tell Stata that A=1 and B=2, so on and so forth. The
> wave variables in the original data set are frivolous anyway - the
> SMOK variables being missing whenever the person failed to respond in
> that wave.
> On 4/26/06, Nick Cox <[email protected]> wrote:
> > I don't think so; but I can't prove that
> >
> > But turn it round. What are you expecting?
> >
> > That -reshape- should
> > somehow realise that A matches 1, B matches
> > 2, C matches 3?
> >
> > What would you expect -reshape- to do if the
> > variables were
> >
> > ID WaveX WaveY WaveZ ASMOK BSMOK CSMOK
> >
> > That one works, but not in the same way.
> >
> > Nick
> > [email protected]
> >
> > > -----Original Message-----
> > > From: [email protected]
> > > [mailto:[email protected]]On Behalf Of Ada Ma
> > >
> > > Carol Bao's friend data set looks interesting. I had not
> noticed the
> > > atwl option before. If the variable names were:
> > > ID WaveA WaveB WaveC ASMOK BSMOK CSMOK
> > >
> > > I could have reshaped it long in one shot:
> > > reshape long @SMOK Wave, i(ID) j(type) atwl() string
> > >
> > > But unfortunate they are not, they are:
> > > ID Wave1 Wave2 Wave3 ASMOK BSMOK CSMOK
> > >
> > > Is it still possible to reshape both the Wave variables
> and the SMOK
> > > variables in one shot?
> > >
> > >
> > >
> > > On 4/26/06, Michael Blasnik <[email protected]> wrote:
> > > > Check out the advanced syntax for reshape. I believe that
> > > you can specify
> > > > the location of the "j" part of the variable names using
> > > the @ symbol. I
> > > > haven't used this option in a while, but I think you will
> > > need to specify
> > > > that the j variable is a string and may need to use the
> > > atwl() option for
> > > > the reshape xij line.
> > > >
> > > > reshape xij @SMOK
> > > >
> > > > Michael Blasnik
> > > > [email protected]
> > > >
> > > > ----- Original Message -----
> > > > From: "Carol Bao" <[email protected]>
> > > > To: <[email protected]>
> > > > Cc: <[email protected]>
> > > > Sent: Tuesday, April 25, 2006 7:01 PM
> > > > Subject: st: reshape when the stub is at the beginning of
> > > the variables
> > > >
> > > >
> > > > > Hi, I am asking this for a friend:
> > > > >
> > > > > She has a data set in its wide form like this:
> > > > >
> > > > > ID Wave1 Wave2 Wave3 ASMOK BSMOK CSMOK
> > > > > 1 0 1 1 . 1 0
> > > > > 2 1 1 0 1 0 .
> > > > >
> > > > > Which reads: for person 1, he/she appears in wave 2
> > > > > and wave 3 of the survey and reports to smoke in wave
> > > > > 2 (BSMOK) and not smoke in wave 3 (CSMOK). For person
> > > > > 2, he/she is interviewed in wave 1 and 2 but drops out
> > > > > in wave 3 and reports to smoke in wave 1 and not in
> > > > > wave 2.
> > > > >
> > > > > The question is, the stub for reshape into variable
> > > > > SMOK appears at the beginning of the variables as A, B
> > > > > and C respectively for three waves (all the other
> > > > > variables as well). Is there any easier way to
> > > > > reshape the data? Or any suggestions?
*
* 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/
