Thanks to Kit Baum, a new command called alignedpairs is now available on SSC. It performs two analyses for matched-pairs: (1) the standard Wilcoxon signed-ranks test using the official Stata signrank, and (2) the Hodges-Lehmann treatment effects estimator with distribution-free confidence intervals. In the case of matched-pairs, the Hodges-Lehmann estimator is defined as the median of the set of n(n+1)/2 Walsh averages. More specifically, the process entails estimating the average difference in outcomes (x-y) for every possible n(n+1)/2 pair and then deriving the overall median of all averages (the Hodges-Lehmann estimator). A distribution-free confidence interval is estimated using large-sample approximation.
For those interested in other Hodges-Lehmann tests, see alignedranks, for the Hodges-Lehmann aligned ranks test on unmatched data, and alignedsets for the Hodges-Lehmann aligned ranks test for matched sets (both available for download at SSC).
As always, I welcome users to share comments and suggestions.
Ariel
For those interested in other Hodges-Lehmann tests, see alignedranks, for the Hodges-Lehmann aligned ranks test on unmatched data, and alignedsets for the Hodges-Lehmann aligned ranks test for matched sets (both available for download at SSC).
As always, I welcome users to share comments and suggestions.
Ariel
