alr4 (version 1.0.0)

Aovtest: Artificial unbalanced anova data.

Description

Artificial data with three factors and a covariate. Used as a homework problem in Weisberg (in press) to explore the various types of analysis of variance.

Usage

data(Aovtest)

Details

The data were generated from a 2 by 3 by 4 design with five observations per cell and one covariate. Use good to select about 80% of the data to set up an unbalanced data set.

References

Weisberg, S. (in press). Applied Linear Regression, 4rd edition. Hoboken: Wiley.

Examples

m1 <- lm(y ~ x + (a + b + c)^3, Aovtest) # all data
m2 <- update(m1, subset=good)  # unbalanced version
m3 <- update(m1, ~ .- x) # orthogonal version, ignore the covariate