eta2.contr: Compute (partial) eta-squared for contrast

In mcfanda/cpower: Power analysis for contrasts

Description

Return the proportion of variance explained by the contrast over the variance not explained by other effects. The result is the partial eta-squared unless, there's only one possible contrast in the model, the is the eta-squared

Usage

eta2.contr(cont, xname, model)

Arguments

cont	vector of contrast weights.
xname	Character string indicating the name of the factor defining groups.
model	a lm object with xname as independent variable.

Details

Compute the eta-squared for a contrast based on a model already estimated with the effect of "xname" af a factor.

Value

A numeric value for eta-squared.

Author(s)

Marcello Gallucci, [e-mail]

See Also

eta2.contr.d

Examples

### sim some data
cont<-c(-3,-1,1,3)
means<-c(10,12,10,12)
y<-rep(means,1000)+rnorm(4000,0,2)
x<-factor(rep(1:4,1000))
### compute eta-squared
mod<-lm(y~x)
eta2.contr(cont,xname = "x",model = mod)

#### check from the d coefficient
d<-d.contr(cont,x=x,y=y)
eta2.contr.d(cont = cont,d=d,scale = "g")

mcfanda/cpower documentation built on May 28, 2019, 1 p.m.