Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: -pexp- and -qexp- revised on SSC

From   "Nick Cox" <[email protected]>
To   <[email protected]>
Subject   st: -pexp- and -qexp- revised on SSC
Date   Fri, 31 Oct 2003 11:22:57 -0000

Thanks to Kit Baum, the packages -pexp- and -qexp- 
have been revised on SSC. 

. ssc inst pexp
. ssc inst qexp 

will install in an up-to-date Stata with internet 

The official Stata commands -pnorm-, -qnorm-, -pchi- 
and -qchi-, on which see the help for -diagplots-, 
provide context. Stata Corp implemented P-P plots
and Q-Q plots for variables against normal (Gaussian, 
central) and chi-square distributions, and then
stopped. Naturally there are many other univariate 
distributions of possible interest for P-P and 
Q-Q plots, and a search will reveal user-written 
commands for several others, including most of the 
usual suspects. 

In revising -pexp- and -qexp- from programs 
written in 1998, what was done was

1 To revise the graphics to Stata 8. In fact, I 
started from the code for -pnorm- and -qnorm- 
and made changes from there. 

2 To support weights, other than pweights. 

3 To allow generation of variables to hold
what was plotted. 

4 To allow fitting of the whole distribution, 
but a graph showing only part of the data, e.g. 
one tail. (This last arose out of a suggestion 
on a different problem by Stephen Jenkins.) 

5 To allow specification of a parameter value (the mean)
as well as fitting the distribution on the fly. 
(In fact this was in the original version.) 

As implied, -pexp- and -qexp- now require Stata 8. 
The original versions, now renamed -pexp5- and
-qexp5-, remain in the package for any interested 
users of Stata 5 through 7. However, they are 
otherwise frozen as was and were not revised 
to incorporate 1 to 4 above. 

[email protected] 

P.S. a design which raises the prospect of 


is in one clear sense poor: who wants so many 
different commands for so many different distributions? 
At some point, I suggest, we need a more general design 
for all the common distributions together with 
an easy way to bolt on distribution-specific code 
for anything else. Perhaps Stata Corp will get 
to this before a user-programmer, or vice versa. 

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index