On Wed, 18 Nov 2009, Henrique wrote:
Please, could you explain me what "non-zero ties" (line
This is a bit obscure. It arises in the context of non-parametric
tests that are based upon rankings. A "tie" is a case where two
or more observations have the same value, and therefore share the
same rank. A "non-zero tie" would be be a case where two or more
observations share a non-zero value.
Specifically, this comes up in the context of the Wilcoxon
signed-rank test. This comment appears in gretl's nonparam.c:
Wilcoxon signed-rank test, with handling of zero-differences
and non-zero ties, plus continuity correction, as in E. Cureton,
"The Normal Approximation to the Signed-Rank Sampling Distribution
when Zero Differences are Present", JASA(62), 1967, 1068-1069.
Frankly, I though this was too obscure to mark for translation,
but the Czech translators noted the non-translated strings and
requested that they be marked.