iquantiles updated on SSC

[Nick Cox]

The command iquantile has been updated on SSC.

The original posting announcing this command is now at https://www.stata.com/statalist/arch.../msg00728.html. The thread was started by
Taggert J. Brooks, whose posting is now at https://www.stata.com/statalist/arch.../msg00652.html

The problem tackled is best explained by example. Suppose you have an ordinal variable with values 1, 2, 3, 4, 5. The median will be reported by most Stata commands as either one of those integers or at most a half-integer, which is coarse and doesn't really use all the information in most examples. On the other hand, many researchers might feel queasy at working with the mean. A compromise is to use the frequencies to interpolate a median.

If that sounds slightly interesting or useful, then use

ssc type iquantile.sthlp

to read as much of the help as you need to help you decide whether to download..

I am aware of an anguished and even angry literature on what summary measures make sense with ordinal scale data, and independently of that the command may have small uses for other kinds of data.

The revision was prompted by Cees van der Eijk, as explained and acknowledged gratefully in the help.

[ericmelse]

Dear Nick,

I tried to acquire the update from ssc, but, when I check the version I get:

which iquantile
*! NJC 1.0.0 19 January 2009

And Cees van der Eijk is not mentioned in the help file.
Possibly something goes awry with iquantile.

[Nick Cox]

ericmelse Thanks for reporting this. KIt Baum reports a glitch at the SSC end -- which should now be fixed.

(Too late now, but the thread title should refer to iquantile not iquantiles.)