Multiple Comparisons (Dunn, Dunn-Bonferroni, & Nemenyi Tests)

Exercise:
  Improvement levels of symptoms of a certain disease by treatment of a pharmaceutical agent with 3 different doses (20mg, 10mg, and 5mg) was assessed. The improvement of symptoms was assessed by various parameters, and expressed as 5 different levels as ‘significantly improved’, ‘moderately improved’, ‘mildly improved’, ‘no change’, and ‘exacerbated’.  

   1) Analyze whether there are significant differences.
   2) When there are significant differences among doses, analyze which doses have significant differences by multiple comparison analysis.

Comparative Verification Test

Kruskal-Wallis test: StatFinale, 0.00303; SPSS, 0.003; EZR, 0.00303.
Multiple comparisons: Exact match with SPSS
Multiple C (Dunn-Bonferroni test): StatFinale & SPSS
Pairwise Comparisons: EZR
Multiple C (Steel test, control vs others): EZR

Reports of SPSS & EZR

SPSS and EZR do not have the tabular format for data entry that Statfinale uses.

※Click pics to enlarge.