Split-split block design: using Location as main plot factor

[Shahal]

Hi, I designed an multi-location experiment.

where, Main plot = Location

sub-plot=Fertilizer (P) source, sub-sub plot = fertilizer rate

My question is,

Should I use location as main plot? Please explain your answer.

SA

[xuelin0820]

Yes, you can if you think location is a fixed effect. If location is a random effect, I would think it is a split-plot design at multiple random locations, then you cannot test the p-value of location, and the random statement should not include Loc*subplot, Loc*mainplot*subplot, because that will give the wrong error term of F test for subplot.