chi-squared statistic for incomplete categorical data
DESCRIPTION:
Calculates the observed-data chi-squared statistic (versus the
saturated model) for incomplete categorical data under a specified
value of the underlying cell probabilities, e.g. one row of the
paramIter component of a
missmodel object.
USAGE:
chi.Loglin(object, theta)
REQUIRED ARGUMENTS:
object
an object of class
"missmodel" whose
paramIter component is of
class
"Loglin". I.e., an object resulting from the use of the
mdLoglin,
daLoglin, or
emLoglin functions. Alternatively,
a class
"preLoglin" object.
OPTIONAL ARGUMENTS:
theta
the model cell probabilities. If not specified, these are obtained as
the final estimates (the last row) in the
"missmodel" object.
VALUE:
the value of the observed-data chi-squared statistic at
theta.
DETAILS:
This is the chi-squared statistic that ignores the missing-data
mechanism.