fit.contrast: Compute and test arbitrary contrasts for regression objects
Description
Compute and test arbitrary contrasts for regression objects.
Usage
fit.contrast(model, varname, coeff, ... )
## S3 method for class 'lm':
fit.contrast(model, varname, coeff, showall=FALSE,
conf.int=NULL, df=FALSE, ...)
## S3 method for class 'lme':
fit.contrast(model, varname, coeff, showall=FALSE,
conf.int=NULL, df=FALSE, ...)
## S3 method for class 'mer':
fit.contrast(model, varname, coeff, showall=FALSE,
conf.int=NULL, sim.mer = TRUE, n.sim = 1000, ...)
Arguments
model
regression (lm,glm,aov,lme) object for which the
contrast(s) will be computed.
varname
variable name
coeff
vector or matrix specifying contrasts (one per row).
showall
return all regression coefficients. If TRUE, all
model cofficients will be returned. If FALSE (the default),
only the coefficients corresponding to the specified contrast will
be returned.
conf.int
numeric value on (0,1) or NULL. If a numeric value is
specified, confidence intervals with nominal coverage probability
conf.int will be computed. If NULL, confidence
intervals will not be computed.
df
boolean indicating whether to return a column containing the
degrees of freedom.
...
optional arguments provided by methods.
sim.mer
Logical value. If TRUE p-values and confidence
intervals will be estimated using [Matrix]{mcmcsamp}. This option only takes effect for mer
objects.n.sim{Number of samples to use in [Matrix]{mcm