R help - Sep 2011 - Data simulation for ANOVA decomposition into sums of squares

Dear listserv,

Please consider the following dataset:

x <- matrix(nrow = 8, ncol = 2)
colnames(matrix) <- c("classification", "soluble_fiber")
x[1:4,1] <- "bagel"
x[5:8,1] <- "donut"

How would I simulate a dataset for a one-way fixed-effect ANOVA (where
"classification" is the treatment variable and
"soluble_fiber" is the response variable) such that the total sums of
squares are equal to 1, and the treatment sums of squares are equal to 0.1? In
other words, I simply want to make up "soluble_fiber" values for each
observation (rows) such that an ANOVA of "soluable_fiber" ~
"classification" yields total sums of squares = 1 and treatment sums
of squares = 0.1.


More generally, I'd like to develop some code that will let me simulate
datasets with different treatment sums of squares (ranging from 0.1 - 0.9) while
holding the total sums of squares constant (equal to 1).

Many thanks in advance for your suggestions. I recognize that this is as much a
statistical/mathematical question as an R programing question.

Sincerely,

-----------------------------------
Josh Banta, Ph.D
Assistant Professor
Department of Biology
The University of Texas at Tyler
Tyler, TX 75799
Tel: (903) 565-5655
http://plantevolutionaryecology.org

	[[alternative HTML version deleted]]