In statistics, when you have an significant interaction between two variables (for exemple concentration and variety). You run an ANOVA test and concentration x variety is significant. How to compare for exemple individually variety 1 and variety 2 for the same concentration ?
One common approach is to look at the simple main effects of one variable at each level of the other--e.g., you may wish to look at the effect of concentration within each variety. If you tell us what software you are using, someone may be able to give more specific advice on how to do that. If you use SPSS, for example, you can obtain tests for the simple main effects via the /EMMEANS sub-command for the UNIANOVA command. (See the link below.)
HTH.
http://www-01.ibm.com/support/docview.wss?uid=swg21475404
You have to estimate Critical difference (CD). the formula is
CD = (Sqrt (EMSS*2/ r))*tat 5% and error df.
EMSS = Error mean sum of square
r = Number of observations on which mean (to be compared in your case) is estimated.
Regards
Hi Maher,
you can use post-hoc tests to compare the simple means, for example S-N-K or Tukey (and maybe also the CD, but I do not know this one... there are so many of them).
I would recommend the Dunn's Multiple Comparison Procedure, because there you can determine how many paired comparisons you want to conduct. This means the test computes less comparisons, it has to correct less for alpha error accumulation and so it becomes more liberal.
Best BS
One common approach is to look at the simple main effects of one variable at each level of the other--e.g., you may wish to look at the effect of concentration within each variety. If you tell us what software you are using, someone may be able to give more specific advice on how to do that. If you use SPSS, for example, you can obtain tests for the simple main effects via the /EMMEANS sub-command for the UNIANOVA command. (See the link below.)
HTH.
http://www-01.ibm.com/support/docview.wss?uid=swg21475404
With a two factor ANOVA, carried out according to the experimental design adopted, you will get THREE (depending on the design, say an RBCD or CRD) LSDs - one for Factor A, one for Factor B, and one for the interaction between the two factors. The LSD for the interaction is what you will need to use to make the comparisons you are referring to.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA