I’m doing a statistics portfolio and one of the questions is asking me to summarize an interaction between 2 independent variables and it says use the appropriate descriptive information to support your interpretation (eg means and/or mean differences).
With a three-factor ANOVA, you need to work on decomposing the effects. The best approach would be two graph each combination of factors or even generate separate graphs depending on exactly what effect is present.
For example, if you have Factor A, B, and C... if the interaction is between Factor A and B with no interaction involving C, generate a bar chart with Factor A as the X-axis. Then, "cluster" Factor B so that it shows a different color for each bar.
You'd repeat the same thing for each interaction. Now, for main effects, you could effectively separate out the different conditions then generate bar charts for the other variables. That is, split the data for Factor A (Group 1A and Group 2A separately). Then generate a graph for Factor B... you'll get one graph for Group 1A's scores with Factor B and a separate for Group 2A's scores with Factor B.
Unpacking multi-factor ANOVA takes quite a bit of work. I hope this helps!
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation