Two and More Factors ANOVA - Agricultural Statistical Data Analysis Using Stata - page 106

In addition to looking at the variety ANOVA tables individually,
the sowing dates also can be examined for each variety with the fol-
lowing command:
contrast date@variety
This results in the following output:
Contrasts of marginal linear predictions
Margins : asbalanced
------------------------------------------------
| df F P>F
-------------+----------------------------------
date@variety |
1 | 2 4.33 0.0214
2 | 2 42.43 0.0000
3 | 2 2.46 0.1012
4 | 2 3.87 0.0310
Joint | 8 13.27 0.0000
|
Residual | 33
------------------------------------------------
With the exception of variety 3 (Nirvana), the other varieties have
significantly lower seedstem numbers with later sowing dates at the p
= 0.05 level of significance. You may wish to try calculating the indi-
vidual ANOVAs to see how they differ from these results.
Split-Plot Design
A split-plot design is another type of factorial design usually used
because of some limitation in space or to facilitate treatment appli-
cation. The two factors are divided into a main-plot effect and a
subplot effect. The precision is greater for the subplot factor than
it is for the main-plot factor. If one factor is more important to the
researcher and if the experiment can facilitate it, then the subplot
factor should be used for this factor. This may not always be the case,
however.