two-way ANOVA

[diamanteskies]

Hi all. I'm wondering why the plots are different when you run with interaction effects and without interaction for a 2-way...
The plot without interaction looks really strange because there seems to be absolutely no variation between the different groups...

[omega]

If plots are different, it suggest that the interaction effect may be significant. You can check the ANOVA results to see if it is true. The plot without the interaction may be ignored.

[diamanteskies]

So if the interaction is not statistically significant, then one should refer to the plot without interaction? but why should the two look different if you are essentially plotting the same things?