I am less clear now on what your real problem is.
But being interested in several variables doesn't
make the issues any different. If you
want to put correlations into variables, you
must still operate with one correlation from
each pair of variables. -matrix accum- looks to me
like a poor choice of method. You end up
putting stuff into a matrix and then need
to get it out again, a bit like going to Houston
from Dallas via San Antonio, perhaps instructive
but nevertheless indirect. As you are cycling
over years and months, you need to work with
several matrices, which makes the problem yet
more complicated.
If I had a varlist a b c d e, there are
ten distinct correlations which we can denote
ab ac ad ae bc bd be cd ce de.
So I can go
foreach v in ab ac ad ae bc bd be cd ce de {
gen `v' = .
}
and then some code to get all the correlations
from all pairs of variables, all years and months
might look like this, but this isn't tested:
tokenize "a b c d e"
qui forval i = 1/4 {
forval j = 2/5 {
forval y = 1988/2003 {
forval m = 1/12 {
corr ``i'' ``j'' if year == `y' & month == `m'
replace ``i''``j'' = r(rho) if year == `y' & month == `m'
}
}
}
}
As you say nothing about what went wrong
as far as you are concerned in your previous
code, I am not sure what you want me to explain.
I can't see an obvious syntax error.
As the FAQ advises, "Say exactly what you typed and exactly what
Stata typed (or did) in response."
Nick
[email protected]
Wang,Le
> I am sorry that I didn't describe my problem properly. I just
> made up the data set.
> There are actually five variables "a, b, c, d, e" from Jan
> 1988 to Dec 2003.
> Although I am aware that there is a -egen corr- availabe, I
> just want to pracitise
> it a little bit since I am not quite familiar with matrix operation in
> STATA. Woud you
> please tell me where I am wrong when I am using matrix form
> in my second post?
> ----- Original Message -----
> From: "Nick Cox" <[email protected]>
> To: <[email protected]>
> Sent: Monday, October 18, 2004 3:10 AM
> Subject: st: RE: loop within loop
>
>
> > If you want the correlation between a and b, what
> > c d e are doing is unnecessary information. A more
> > direct approach is
> >
> > gen corrab = .
> > qui forval i = 1988(1)2003 {
> > forval j = 1/12
> > corr a b if year == `i' & month == `j'
> > replace corrab = r(rho) if year == `i' & month == `j'
> > }
> > }
> >
> > In your example, each correlation is based on just 3 data points.
> > Perhaps you are obliged to do that, but I wouldn't want to.
> >
> > Also, note Nick Winter's -egen, corr()- within -egenmore- on SSC.
*
* 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/
