I would like to generate a matrix to plan an experiment.
- I have 6 individuals and 6 treatments delivered in 6 days.
- I can use 1 individual per week (6 weeks of 6 days in total = 36 days of testings with 6 treatments a day).
- Every day, I will test 6 treatments on the individual in a random order. An individual should not start with the same treatment any day of the week.
For example:
Monday: A C D B E F
Tuesday: C E A B F D
...
- If possible, different individuals should not have the same order that the others.
- I would like to take into account order treatment; that is each pair of treatments following each other should be repeated as equally often as possible to avoid "carryover" (sequence and order) effects. I am working with R at the moment but could also work with MATLAB.
I have tried to work out a design with the package "agricolae" and the package "crossdes" but no success so far. My ideas were to use a balanced incomplete block design. An example of one of my code follows.
> trt <- c("A", "B", "C", "D", "E", "F" )
> k <- 6
> outdesign<- design.bib (trt, k, seed=0)
Error in md[i, bi] : incorrect number of dimensions