Hi all. Any help will be greatly appreciated.
I am trying to decide which post hoc test is the correct one to use to find the differences between groups after I un an ANOVA.
I started with the Scheffe test because my group sample sizes are different (varying from 20 to 60 cases). When I run an ANOVA I get a sig value of less than .05 – a significant statistical difference. I was expecting when I ran the Scheffe post hoc I would be able to see the differences between groups because the ANOVA sig value is less than 0.5. In the Multiple Comparisons table I saw no difference.
When I ran the Tukey and Tukey b I was able to see the sig differences (value of less than .05 ) between groups in the Multiple Comparisons table.
My questions are:
1. When using the Scheffe test why did I not see the sig differences in the Multiple Comparisons table when the ANOVA table shows there are? Is this due to the discordant sample size or something else?
2. Why did I see the sig differences when I ran the Tukey and Tukey b tests?
3. Which of the three test should I use or should I use a combo?