Factor means
Consider a split plot experiment with factor A allocated at block level and factor B varied within blocks. If factor B's levels are used in the same proportion within all blocks, tables of factor means (or graphical display of these means) can be used to summarise the effects of the factors.
Maize seed planting
To illustrate, the diagram below summarises the main effects of seedbed preparation and planting method.
The p-value for testing the interaction was 0.1988, so there is no evidence of an interaction. The overall mean responses for each method of seedbed preparation and for each planting method are initially shown in the diagram and provide good summaries of the main effects of the two factors. It is clear that planting method B2 gives much lower yields than the other planting methods but all other planting methods and all methods of seedbed preparation are similar.
Click Show interaction. To summarise the interaction between the two factors, the mean yields for each combination of seedbed preparation and planting method must be displayed. The diagram shows these 16 means in two different ways — against horizontal axes for the separate factors.