Could anyone tell me why in a two-way analysis of variance (ANOVA) used to examine the main and interaction effects of different m parameter (m = 2, 4, 6, 8, 10) and signal type (WHOLE, SEGM, NORM, ZERO) in HSD Tukey post-hoc did not show the differences in m=8 and 10 between the Zero and Segm types (marked on the figure)??? vertical bars on the figure are the confidence intervals. could it be due to breaking the assumption of homogeneity of variance? If yes, one possible reason of breaking this assumption could be that mean values are correlated with SD? but if that was the case, the confidence intervals would also be bigger and here CI do not change / increase despite higher means ... Anyone know other reasons?
Dear Anna,
Did you check the normality test? if yes, is it non-significant?
My regards
Actually it does show Statisically significantly differences anyplace the CIs don't overlap. Since I don't know what the experiment was or the research questions were. I have no idea what any of that means. Please remember statistics is not magic Best wishes, David Booth
Ahmad Alshallawi in most cases the distribution was normal, besides according to the literature ANOVA suppose to be resistant for violation of this assumption.
Anna Brachman. In this study, it was stated whether the Anatolian Diagonal created a difference in the wing characteristics of O.brunneum insects within the species. As David Eugene Booth stated, there is no difference in some points, we can think of it as a difference in other points.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA