Re: st: Problem with statsby and correlate

[n j cox <n.j.cox@durham.ac.uk>]

As you report, -correlate- only leaves behind one covariance
in memory. In any case, you can't stuff a matrix into
a single value of a single observation, which in effect
you are trying to do.

You need a very different approach. Off-hand,
I am not clear what is best for your data structure.
It may well  be that you need a different data structure.
If this were my problem, I would -reshape long- first.

Nick
n.j.cox@durham.ac.uk

Pavlopoulos.D.

I want to produce a file with the autocovariance matrix of wage for 7 
subsequent years. I have my data organized in such a way that I have one 
income variable per year, so the wage variables in my dataset are:

lhwage1995 lhwage1996 lhwage1997 lhwage1998 lhwage1999 lhwage2000 lhwage2001


Then, I run:

statsby ,by(country) clear: corr lhwage1995 lhwage1996 lhwage1997 
lhwage1998 lhwage1999 lhwage2000 lhwage2001, covariance

But then I get a dataset where only the covariance between lhwage1995 
lhwage1996 for every country is saved. The rest of the covariances 
disappear. Do you know how can I solve this? The `by' option is not the 
one creating the problem. I tried to remove it but things didn't change.

*
*   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/