Bolstad (version 0.2-41)

xdesign: Monte Carlo study of randomized and blocked designs

Description

Simulates completely randomized design and randomized block designs from a population of experimental units with underlying response values \(y\) and underlying other variable values \(x\) (possibly lurking)

Usage

xdesign(
  x = NULL,
  y = NULL,
  corr = 0.8,
  size = 20,
  n.treatments = 4,
  n.rep = 500,
  ...
)

Arguments

x

a set of lurking values which are correlated with the response

y

a set of response values

corr

the correlation between the response and lurking variable

size

the size of the treatment groups

n.treatments

the number of treatments

n.rep

the number of Monte Carlo replicates

…

additional parameters which are passed to Bolstad.control

Value

If the ouput of xdesign is assigned to a variable, then a list is returned with the following components:

block.means

a vector of the means of the lurking variable from each replicate of the simulation stored by treatment number within replicate number

treat.means

a vector of the means of the response variable from each replicate of the simulation stored by treatment number within replicate number

ind

a vector containing the treatment group numbers. Note that there will be twice as many group numbers as there are treatments corresponding to the simulations done using a completely randomized design and the simulations done using a randomized block design

Examples

# NOT RUN {
# Carry out simulations using the default parameters 

xdesign()

# Carry out simulations using a simulated response with 5 treaments, 
# groups of size 25, and a correlation of -0.6 between the response 
# and lurking variable

xdesign(corr = -0.6, size = 25, n.treatments = 5)

# }