Interaction term in two-way anova

#1
Dear all,

I would like to see the effect of two independent variable (a,b) on a dependant variable (c). a and b are covariate.

Could you please let me know how to do two-way anova if

1) there is an interaction between a and b
2) there is no interaction between a and b

I am quite confused with this.

I've done the two-way anova with and without including the interaction term (a*b), the results are totally different.

When I include the interaction term, a and b significantly affect c (both p values are less than 0.01), but when I don't include the interaction term, a affects c but b doesn't (p value greater than 0.01).

Thank you,
Aj
 
#2
First: If the interaction makes the other results different, then you should include it.
Second: When there is an interaction, the meaning of the main effects changes. Indeed, an interaction means that the effect of one IV on the DV is different at different levels of the other IV. So, with an interaction, you can't say "a affects c" or "a does not affect c" you can only say what the effect of a on c is at each level of b. Similarly, you can't say "b affects c" or "b does not affect c".