Sets up design matrix for measurement error models.
me(formula, error=NULL, group=NULL, type="classical")
design matrix, with a prior distribution attribute
formula
for the fixed effects.
character; name of column in data.frame
in which standard error (type="classical"
or type="berkson"
) or miscalssification error (type="dclassical"
) is stored.
name of column in data.frame
in which groups are stored. Rows of the design matrix with the same group
level are assumed to pertain to the same obsevation of the covariate that is measured with error.
character; one of type="classical"
, type="berkson"
, type="dclassical"
or type="dberkson"
(see details)
Jarrod Hadfield j.hadfield@ed.ac.uk