How it works...

The ANOVA table shows the degrees of freedom, the sum of squares, the mean squares (sum of squares divided by the degrees of freedom), the F statistic value, and the p-values. As we stated previously, we need to focus our attention on the interaction: it can be seen that it is not significative at the 0.05 level. Once it is rejected, we can proceed with our analysis: both the color and the font are highly significative. We can then conclude that Colour and the Font type explain a substantial part of the variability. This implies that, for both effects, at least one of the levels (one of the colors or one of the font types) has a mean that is different from the rest. Some statisticians prefer to re-estimate the model without the interaction in case it has been found to be non-significative.