factDesign (version 1.48.0)

contrasts: Construct appropriate lambda matrix and test linear contrasts of the parameter estimates from a linear model.

Description

'par2lambda' takes list of lm coefficient names and a corresponding list of numeric vectors corresponding to hypothesis tests of linear contrasts and returns a lambda matrix suitable for an F-test of the linear contrasts. 'par2lambda' is intended to be used in conjunction with 'contrastTest' or 'findFC'.

'contrastTest' performs an F test for simultaneous tests of linear contrasts using an appropriately specified lambda matrix and an lm object.

Usage

par2lambda(betaNames, betas, coefs)
contrastTest(model, lambda, cVec = NA, p = 0.01)

Arguments

betaNames

A character vector of the names of the coefficients in a linear model.

betas

A list of vectors of the parameters to be used in the contrasts.

coefs

A list of vectors of numeric coefficients corresponding to betas.

model

An lm object.

lambda

A matrix of coefficients in the appropriate order to be multiplied by the estimated coefficients of the lm object, possibly returned from 'par2lambda'.

cVec

A vector of constants for testing that the linear contrasts equal something other than zero. If this is unspecified, it is assumed to be zero.

p

The significance level at which to perform the contrast test.

Value

For par2lambda: A lambda matrix constructed for testing linear contrasts using lm output.For contrastTest:

test

Returns "REJECT" or "FAIL TO REJECT" based on the result of the test of hypothesis.

Fstat

The F statistic for the test of contrast.

pvalue

The corresponding pvalue for the F test.

cEst

The contrast estimate.

See Also

lm

Examples

data(estrogen)
ES <- pData(estrogen)[["ES"]]
TIME <- pData(estrogen)[["TIME"]]   
fit <- lm(exprs(estrogen)["40079_at",] ~ ES + TIME + ES*TIME)
betaNames <- names(coef(fit))
betas <- list(c("ESP"),c("ESP","ESP:TIME48h"))
coefs <- list(c(1),c(1,1))
lambda <- par2lambda(betaNames,betas,coefs)
contrastTest(fit,lambda)