st: -qenv- package posted on SSC

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

Thanks to Kit Baum as usual, a new -qenv- package has been posted on SSC. It contains programs -qenvnormal- and -qenvgamma-. The first of these is a considerable tidying-up of code published as -qnormenv- on 5 and 6 March at 

http://www.stata.com/statalist/archive/2012-03/msg00244.html

in a thread on -qnorm- started in 

http://www.stata.com/statalist/archive/2012-03/msg00150.html

The second of these is new, namely -qenvgamma-. -qenvnormal- requires Stata 9 and -qenvgamma- requires Stata 10.1. Use -ssc- to install if interested. 

The two programs each generate two numeric variables for later plotting on a quantile-quantile plot that define an envelope of low and high expectations for each quantile given repeated sampling from a specified distribution with the same number of values and same parameter values as the data specified. The programs are helper programs designed for use with -qplot- from SJ, which must be installed separately.  (Type -findit qplot- to get references and code sources and please note that an update of -qplot- is itself imminent in Stata Journal 12(1).) 

The publication of separate programs for normal and gamma distributions (as may be guessed from the names) may provoke an expectation that these will be followed in due course by a small flurry of other programs for other distributions, or at least those (continuous) distributions that I care about personally. What is much more likely is that when next revisited there will be an overall program -qenv- to work with various distributions, although it is still likely to be restricted to those distributions I care about, although I will try to think of ways of leaving a back door for people to specify their own favourite distribution so long as they have a way of fitting it and picking up the resulting parameter values. A next generation of -qenv- will, I hope, include an option for users to specify parameter values other than those estimated from the data specified. 

Let me please also make clear that although the thread mentioned above was (initially) about the official command -qnorm- nothing here makes direct or indirect use of that command; and also that the scope of this package is restricted to marginal univariate distributions with nothing said about any set of predictors. 

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



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/