raov(formula, data, ...)
aov
if the appropriate data
frame is attached or the variables are on the search path.
lm
defining the model
or the fit, such as
subset
or
na.action
.
aov
but containing
two additional components used for computing the estimated random effects:
The difference between
raov
and
aov
appears when the summary method
is called;
this method will compute and print the estimated variances for the
random effects model with the output from
raov
.
Specifically, the ordinary mean-squares and the component
ems.coef
from the fitted anova model are used to estimate the variances for
the usual random effects model for each relevant main factor or interaction
in the model.
That is, the so-called ANOVA method is used to estimate the variance components.
The model must be balanced (
raov
checks this with the
replications
function), and the model must be fully random.
Only single strata designs are
currently allowed; i.e., there can be no
Error
term in the formula.
praov <- raov(Moisture ~ Batch/Sample, pigment) summary(praov) Df Sum of Sq Mean Sq Est. Var. Batch 14 1210.93 86.4952 7.1280 Sample %in% Batch 15 869.75 57.9833 28.5333 Residuals 30 27.50 0.9167 0.9167