DD {LabApplStat}

Design diagram for a linear model

Description

DD computes the Design Diagram for a linear model.

Usage

DD(fixed, random = NULL, data, keep = ~1, center = FALSE, eps = 1e-12)

Arguments

fixed

formula with fixed effects. A response may the specified, but this optional.

random

formula with random effects. Defaults to NULL meaning that there are no other random effects than the residual, which is added to all designs.

data

data frame with the explanatory variables and the response (if specified).

keep

formula which effects that will not be removed in the collinarity analysis. Defaults to ~1 meaning that the intercept will be kept if it is present.

center

boolean deciding whether to centralize numerical predictors when an intercept is present. Defaults to FALSE.

eps

threshold for deeming singular values to be "zero". Defaults to 1e-12.

Value

An object of class designDiagram-class

Author(s)

Bo Markussen

See Also

minimum, plot.designDiagram

Examples

# 3-way ANOVA
x <- factor(rep(rep(1:4,times=4),times=4))
y <- factor(rep(rep(1:4,times=4),each=4))
z <- factor(rep(rep(1:4,each=4),each=4))
myDD <- DD(~x*y*z,data=data.frame(x=x,y=y,z=z))
summary(myDD)

#Making the factor diagram closed under minima
mydata <- data.frame(age=rep(c("boy","girl","adult","adult"),4),
                     gender=rep(c("child","child","man","woman"),4))
myDD <- DD(~0+age+gender,data=mydata)
plot(myDD)

# Example of collinearity
mydata <- data.frame(age=rnorm(102),edu=rnorm(102),sex=factor(rep(c(1,2),51)))
mydata <- transform(mydata,exper=age-edu+0.1*rnorm(102))
mydata <- transform(mydata,wage=2*edu+2*exper+rnorm(102))
summary(myDD <- DD(wage~sex*(age+exper+edu),data=mydata))

# growth of rats
antibiotica <- factor(rep(c(0,40),each=6))
vitamin <- factor(rep(rep(c(0,5),each=3),2))
growth <- c(1.30,1.19,1.08,1.26,1.21,1.19,1.05,1.00,1.05,1.52,1.56,1.55)
mydata <- data.frame(antibiotica=antibiotica,vitamin=vitamin,growth=growth)
myDD <- DD(growth~antibiotica*vitamin,data=mydata)
plot(myDD,"MSS")
plot(myDD,"I2")


  # ANCOVA: Non-orthogonal design
  library(isdals)
  data(birthweight)
  plot(DD(weight~sex*I(age-42),data=birthweight),"MSS")
  plot(DD(weight~I(age-42)+sex:I(age-42)+sex,data=birthweight),"MSS")