0

I am using the package PMCMR to perform a postdoc Kruskal-Nemenyi test. When I run the test with the default settings:

posthoc.kruskal.nemenyi.test(x=coastal$HIGH_MORTGAGE, g=coastal$SIZECLASS, method="Tukey")

I get the following results and warning:

Pairwise comparisons using Tukey and Kramer (Nemenyi) test  
                   with Tukey-Dist approximation for independent samples

data:  coastal$HIGH_MORTGAGE and coastal$SIZECLASS 

       Large Medium
Medium 0.931 -     
Small  0.746 0.078 

P value adjustment method: none

Warning message:
In posthoc.kruskal.nemenyi.test.default(x = coastal$HIGH_MORTGAGE,  :
  Ties are present, p-values are not corrected.*

When I run the test changing the distribution to Chisq to apply a correction for the ties, I still get the same results and the chi-square distribution is not used.

posthoc.kruskal.nemenyi.test(x=coastal$HIGH_MORTGAGE, g=coastal$SIZECLASS, method="Chisq")

Pairwise comparisons using Tukey and Kramer (Nemenyi) test  
                   with Tukey-Dist approximation for independent samples

data:  coastal$HIGH_MORTGAGE and coastal$SIZECLASS

       Large Medium
Medium 0.931 -     
Small  0.746 0.078

P value adjustment method: none 

Warning message:
In posthoc.kruskal.nemenyi.test.default(x = coastal$HIGH_MORTGAGE,  :
  Ties are present, p-values are not corrected.

I would like to know if there is a bug in the package or if there is any methos I am not aware of to solve this.

Cecilia
  • 1
  • 1
  • 2

1 Answers1

1

From version PMCMR1.0 to PMCMR1.1 (and > 1.1) the syntax has slightly changed, so it is dist, not method, it is:

posthoc.kruskal.nemenyi.test( x, g, dist = c("Tukey", "Chisquare"), ...)

or

posthoc.kruskal.nemenyi.test(formula, data, subset, na.action, dist =
 c("Tukey", "Chisquare"), ...)

The vignette that has included an example has been updated in version PMCMR1.3 accordingly.

Dimitris Fasarakis Hilliard
  • 150,925
  • 31
  • 268
  • 253
Thorsten Pohlert
  • 11
  • 1