0

I ran an oneway ANOVA with 9 steps. In the options I clickt the LSD post-hoc test. In Andy Fields Statistics Book there is a hint that LSD requires the overall ANOVA to be sig. But it doesn't tell what I should do if I checked the box for LSD and SPSS gives me that as well in my output. And now there are some sig differences. Can i still report them and later use them for interpretation? Just with the note that the overall wasn't significant?

Thanks for helping me Kerstin

Kerstin
  • 3
  • 1

1 Answers1

0

According to Do we need a global test before post hoc tests? you cannot proceed with the LSD test if the global F test does not reject the null hypothesis. The purpose of checking the global F test first is to correct for multiple comparisons when performing the LSD test. However, other post hoc tests have the correction built in and therefore do not require an initial check of the global F test.

Joel Galang
  • 138
  • 6