[Vegan-commits] r1424 - in branches/1.17: R inst

[noreply at r-forge.r-project.org]

Author: jarioksa
Date: 2010-12-15 15:52:19 +0100 (Wed, 15 Dec 2010)
New Revision: 1424

Modified:
   branches/1.17/R/mantel.correlog.R
   branches/1.17/inst/ChangeLog
   branches/1.17/inst/NEWS
Log:
merge 1423 to branches: final fix for the release?

Modified: branches/1.17/R/mantel.correlog.R
===================================================================
--- branches/1.17/R/mantel.correlog.R	2010-12-15 07:01:46 UTC (rev 1423)
+++ branches/1.17/R/mantel.correlog.R	2010-12-15 14:52:19 UTC (rev 1424)
@@ -100,7 +100,7 @@
                 ## (H1: r>0) Here, compute a one-tailed p-value in
                 ## direction of the sign
                 if(temp$statistic < 0) {
-                    temp.p <- ((1-temp.p)*(nperm+1))/(nperm+1)
+                    temp.p <- (sum(temp$perm <= temp$statistic)+1)/(nperm+1)
                 }
                 mantel.p <- c(mantel.p, temp.p)
             } else {

Modified: branches/1.17/inst/ChangeLog
===================================================================
--- branches/1.17/inst/ChangeLog	2010-12-15 07:01:46 UTC (rev 1423)
+++ branches/1.17/inst/ChangeLog	2010-12-15 14:52:19 UTC (rev 1424)
@@ -2,7 +2,7 @@
 
 VEGAN RELEASE VERSIONS at http://cran.r-project.org/
 
-Version 1.17-5 (released December 15, 2010)
+Version 1.17-5 (released December 16, 2010)
 
 	* based on development version 1.18-18 at
 	http://vegan.r-forge.r-project.org/ 
@@ -83,6 +83,8 @@
 
 	* merged r1418: Almeida-Neto & Ulrich got published.
 
+	* merged r1423: fix to p-values in mantel.correlog.
+
 	* sipoo.rda got row names and decision-vegan.Rnw changed
 	accordingly in parallel to pkg/vegan devel branch.
 

Modified: branches/1.17/inst/NEWS
===================================================================
--- branches/1.17/inst/NEWS	2010-12-15 07:01:46 UTC (rev 1423)
+++ branches/1.17/inst/NEWS	2010-12-15 14:52:19 UTC (rev 1424)
@@ -36,11 +36,14 @@
       2.12.1. Uses zapsmall() to print near-zero
       eigenvalues.
 
-    - mantel.correlog: P-values were slightly off, because function
-      added 1 to the numerator and denumerator, although this
-      correction was already done in mantel(). (The correction was not
-      done in older version of mantel(), but the mantel.correlog() was
-      not changed after fixing mantel()).
+    - mantel.correlog: P-values could slightly off. The function uses
+      internally mantel() to evaluate the statistic and its
+      significance. The mantel() function adds one both to the
+      denominator and numerator, but mantel.correlog() did not notice
+      this and made the addition for the second time. When the
+      mantel.correlog() reported a positive value of the statistic, it
+      reversed the direction of the one-sided test of mantel(), but
+      did not handle tied values correctly in this reversal.
    
     - nestednodf: Gustavo Carvalho's new version that also implements
       a new quantitative method of Almeida-Neto & Ulrich (Env Mod