[R] Repeated measures on a factorial unbalanced in a blocks with split-plot design

[Marcelo Laia]

Dear all,

I am trying to analyze data from an experiment like this:

Factors:
Hormone - Levels: SH, CH (S = without; C=with; H=Hormone)
Time - Levels: 19/08/09, 04/09/09, 18/09/09, 08/10/09, 20/10/09 (DD/MM/YY)
Nutrition - Levels: Completa, Sem (without)
Macronutrition - Levels: Ca, K, Mg, P, Sem (without)

Time is the measures day. It reflect the days after germination.

Blocks : 4
plants per sub-plots: 16

Each plot was divided in two parts equals. In each part, there was 6
sub-plots with 16 plants (2x8 plants).

In the first part of the plot, it was treated with CH and other-one
was treated with SH. No randomization here.

Factors Nutrition and Macronutrition was combined together:

Treatment 1 - Completa, Sem
Treatment 2 - Completa, Ca
Treatment 3 - Completa, Mg
Treatment 4 - Completa, P
Treatment 5 - Completa, K
Treatment 6 - Sem, Sem (control: without Hormone, without Nutrition,
and without Macronutrition)

This six treatments were randomized on each sub-plot in "CH" and "SH".
Randomization was different for each Block.

However, treatment 6 is not present on CH. It is only present on SH range.

Here was a experimental design:

http://www.divshare.com/download/9241232-392

Each Time, we measured Diametro (centimeters), Altura (centimeters),
and N.Folhas (count).

We are interested on treatments effects on Diameter (Diametro), High
(Altura), and leaves number (N.Folhas). Are there effects? Are there
time effects? And interactions? How is the best time, and the best
nutrition, and the best macronutrition combination? How is the
influence of hormone? And interactions?

I try this approach, but I don't know how I could handle the repeated
measures here! Nor if this approach is correct for me.

Here is a subset of my data

http://www.divshare.com/download/9241231-428

> dados <- read.table("marcelo_laia.txt",sep="\t",dec=",",header=TRUE)
> summary(dados)
> dados.model <- aov(Diametro ~ Block +
+                     Hormone + Error(Block/Hormone) +
+                     Treatment + Treatment:Hormone +
+                     Hormone/Block/Treatment,
+                     data=dados)
> summary(dados.model)

This model was correct? (T6 was present only on SH range)

How I could include the repeated measures here?

Thank you very much!

--
Marcelo Luiz de Laia
Universidade do Estado de Santa Catarina
UDESC - www.cav.udesc.br
Lages - SC - Brazil
Linux user number 487797