One-way ANOVA post host procedure (let's say by using Tukey HSD) in SPSS includes two tables: the multiple comparisons table and the homogeneous subsets. Most of the times these two tables give the same results. However, in some cases the results of the two tables are different. In such cases, which one is the most appropriate to use in order to explain the differences between the means of the groups?
This is very interesting.
An online example of this phenomenon can be seen at the following website at the first instance of Tukey. http://www.uvm.edu/~dhowell/gradstat/psych341/lectures/AnovaReviewFolder/class2.html
Thank you Salvatore. Interesting website. I noticed that the opinions of researchers on this subject are contradictory. Some suggest that in the case of unequal samples it is better to use homogeneous subsets, while others believe that even in that case we have to follow the results of multiple comparisons.
- Badges
- Science topic