The residuals at level \(i\) are obtained by subtracting the fitted values at that level from the response vector. The fitted values at level \(i\) are obtained by adding together the contributions from the estimated fixed effects and the estimated random effects at levels less or equal to \(i\) and evaluating the model function at the resulting estimated parameters.
# S3 method for nlmeStruct
residuals(object, level, conLin, …)an object inheriting from class "nlmeStruct",
representing a list of mixed-effects model components, such as
reStruct, corStruct, and varFunc objects.
an optional integer vector giving the level(s) of grouping
to be used in extracting the residuals from object. Level
values increase from outermost to innermost grouping, with
level zero corresponding to the population fitted values. Defaults to
the highest or innermost level of grouping.
an optional condensed linear model object, consisting of
a list with components "Xy", corresponding to a regression
matrix (X) combined with a response vector (y), and
"logLik", corresponding to the log-likelihood of the
underlying nlme model. Defaults to attr(object, "conLin").
optional arguments to the residuals generic. Not used.
if a single level of grouping is specified in level,
the returned value is a vector with the residuals at the desired
level; else, when multiple grouping levels are specified in
level, the returned object is a matrix with
columns given by the residuals at different levels.
Bates, D.M. and Pinheiro, J.C. (1998) "Computational methods for multilevel models" available in PostScript or PDF formats at http://nlme.stat.wisc.edu