Thanks to Kit Baum, the dstat package has been updated on SSC. To install the update, type

Code:
. ssc install dstat, replace
or use the adoupdate command. Note that dstat requires moremata and coefplot, so possibly also type

Code:
. ssc install moremata, replace
. ssc install coefplot, replace
Main changes:
  • There is a new option nocasewise that allows you to exclude missing values on the basis of individual variables instead of listwise deletion. If nocasewise is specified, all nonmissing observations of a variable will be used in the computations even if there are missing values on other variables included in the call to dstat.
  • There is a new option relax that will avoid error if there are observations outside of the support of a particular statistic (e.g. observations equal to zero in case of the Theil index or the MLD). Computations will be based on all valid observations.
  • Standard error estimation for frequencies and totals has been revised so that it is consistent with the assumptions made by svy and bootstrap. In the previous version, suboption svy had to be specified within vce(analytic) or vce(cluster); this is no longer necessary and the suboption has been removed.
  • some minor bug fixed etc; see the change log at https://github.com/benjann/dstat/
For those who are interested, dstat will be presented at the German Stata Conference on June 25 and the US Stata Conference on August 5/6.

ben