agridat (version 1.12)

weiss.incblock: Soybean balanced incomplete block experiment

Description

Soybean balanced incomplete block experiment

Usage

data(weiss.incblock)

Format

A data frame with 186 observations on the following 5 variables.

block

block factor

gen

genotype (variety) factor

yield

yield (bu/ac)

row

row

col

column

Details

Grown at Ames, Iowa in 1937. Each plot was 6 feet by 16 feet (2 rows, 3 feet apart). Including space between plots, the entire experiment was 252 ft x 96 feet (7 block * 6 plots * 6 feet = 252, 16*5 plots plus 4 gaps of 4 feet). Weiss shows a figure of the field (that was later doubled in dize via using two rows per plot).

Note that only 30 varities were tested. Varieties 7 and 14 are the same variety (Mukden). Although total yields of these varieties were not equal, the correction for blocks adjusted their means to identical values. Such accuracy is not, however, claimed to be a constant characteristic of the design.

Examples

# NOT RUN {
data(weiss.incblock)
dat <- weiss.incblock

# True aspect as shown in Weiss and Cox
desplot(yield~col*row, dat, text=gen, shorten='none', cex=.8, out1=block,
    main="weiss.incblock", xlab="Soybean yields", aspect=252/96)

# }
# NOT RUN {
  # Standard inc block analysis used by Weiss and Cox
  require(asreml)
  m1 <- asreml(yield ~ gen + block , data=dat)
  predict(m1, classify="gen")$predictions$pvals
  ## gen pred.value std.error est.stat
  ## G01     24.59    0.8312  Estimable
  ## G02     26.92    0.8312  Estimable
  ## G03     32.62    0.8312  Estimable
  ## G04     26.97    0.8312  Estimable
  ## G05     26.02    0.8312  Estimable
# }
# NOT RUN {
# }