Thanks to Kit Baum, various packages have been
updated on SSC.
If you installed any of these previously, please
update using -ssc-.
If any sounds possibly interesting or useful,
then please install, again with -ssc-, but note
various limits on Stata version.
-countmatch-
============
Requires Stata 8. This was posted a few days ago
following a thread started by Socrates Mokkas.
It replaces an older, buggy -fndmtch2-.
The key question is: I have two variables -var1- and -var2-.
I want to find values in -var2- equal to those of -var1-,
in either the same or other observations.
An easy example, which I thought of after the Statalist
thread, is that friendship data are stored as -friend-
and -bestfriend-. So you may be "George" and your best
friend may be "Donald". Or perhaps not. Anyway,
a natural question now is, how many people name
each person in -friend- as their best friend in -bestfriend-?
Sports fans will be able to think of similar examples
in terms of home and away teams, political scientists
in terms of aggressors and victims, and so forth.
Marcello Pagano pointed out some awkward and
unclear wording in the help. I addressed that
and also added yet more suggestions to the help
on related problems.
-lmoments-
==========
Requires Stata 9.
This is for L-moments. See the help for details.
Recently I posted a major update incorporating
Mata code written with the help of Bill Gould.
Stephen Jenkins found a bug lurking in a corner
which is now fixed. (Don't worry; if it had bitten
you, you would have known about it.)
-nruns-
=======
Requires Stata 9.
This is for getting probabilities of run numbers
under randomness. It was written in collaboration
with Nigel Smeeton.
Kit Baum and Bill Gould made various suggestions
for tweaking the Mata code that led to about 20%
speed-up.
-transint-
==========
Requires Stata 8.
This is just a long help file on transformations
which I wrote some years ago because I couldn't find
what I wanted in the literature, an explanation of
transformations that is unified and talks about more
than taking logs or square roots, but also remains
fairly elementary in level. The last substantial
revision was a year ago.
There is also some discussion of a problem recently aired,
what can be done when you want something like logs but your
data contain zeros?
On a re-reading, I found several typos, including
some missing minus signs, which I have fixed.
Nick
n.j.cox@durham.ac.uk
*
* 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/