Adding coin package

R/wilcoxonsr_hodges_lehman.qmd

@@ -118,4 +118,38 @@ DOS2::senWilcox(
 -   Gamma \>= 1 is the value of the sensitivity parameter. If gamma=1, then you are assuming ignorable treatment assignment or equivalently no unmeasured confounding - that is the considered scenario in our example, sensitivity analysis is not performed.
 -   Only p value, estimate and CI are provided
 
-### Coin package - coming soon! {.unnumbered}
+#### Coin {.unnumbered}
+
+### Examples {.unnumbered}
+
+```{r}
+# Exact
+# Default method of handling zeros - Pratt (1959):
+coin::wilcoxsign_test(blood_p$bp_after~blood_p$bp_before,
+                zero.method = "Pratt",
+                distribution = "exact")
+
+# Wilcoxon (1949) method of handling zeros
+# the same results for datasets with no ties and no 0s
+coin::wilcoxsign_test(blood_p$bp_after~blood_p$bp_before,
+                zero.method = "Wilcoxon",
+                distribution = "exact")
+
+# Asymptotic approximation & No continuity correction
+coin::wilcoxsign_test(blood_p$bp_after~blood_p$bp_before,
+                zero.method = "Pratt",
+                distribution = "asymptotic")
+
+# Monte Carlo resampling (& No continuity correction)
+coin::wilcoxsign_test(blood_p$bp_after~blood_p$bp_before,
+                zero.method = "Pratt",
+                distribution = "approximate")
+
+
+
+```
+
+### Important notes on coin::wilcoxsign_test {.unnumbered}
+-   Only p value, Statistic V and estimate are provided, no H-L CI is calculated
+-   The default method of handling zeros (Pratt), first rank-transforms the absolute differences (with zeros) and then discards the ranks corresponding to the zero-differences. The alternative one (Wilcoxon) first discards the zero-differences and then rank-transforms the remaining absolute differences
+-   No continuity correction is applied and there is no option to change that