So i did an experiment with 4 groups, each getting different advertisement communications tasty or healthy benefits:
group 1: control group, no health or taste benefits
group 2: only health benefits
group 3: only taste benefits
group 4: health & taste benefits
I have 4 different dependant variables: purchase intention, attitude, credibility and benefits all measured on 7 point likert scale
So I thought i had to do a 2 way anova because health and taste are the IV. But how do i test these? Because the interaction of them is already tested in group 4?
I work with SPSS
Best Answer
The two methods are statistically equivalent, but a two-way ANOVA will make it easier to test the specific effects of interest (i.e., an interaction between health benefits and taste and main effects of each). The model F-test in a two-way ANOVA will be equivalent to the F-test in the one-way ANOVA.
You can use a one-way ANOVA to test main and interaction effects. To test the main effect of health, you can test whether $\frac{\bar{x}_1+\bar{x}_3}{2}=\frac{\bar{x}_2+\bar{x}_4}{2}$. The t-test for this comparison will be equivalent to the F-test on the main effect of health benefits in a two-way ANOVA. For the interaction, you can test whether $\bar{x}_2-\bar{x}_1=\bar{x}_4-\bar{x}_3$. The t-test on that will be equivalent to the interaction F-test in a two-way ANOVA. A one-way ANOVA makes it easier to compare individual groups and other contrasts (e.g., is there a difference between group 1 and all other groups combined?) while a two-way ANOVA makes it easier to test interactions, main effects, and simple main effects.