fitted glm model object.
This is assumed to be the result of some fit that produces
an object inheriting from the class
"glm", in the sense that
the components returned by the
glm() function will be available.
OPTIONAL ARGUMENTS:
dispersion
a supplied value for the dispersion parameter.
The default is
1 for the
binomial or
poisson families.
For other families the default is the residual Chi-square statistic divided
by the residual degrees of freedom.
For the Gaussian family, for example, the dispersion parameter is
the error variance and the default is the Residual sum of
squares divided by the residual degrees of freedom.
Supplying a value of
0 causes the Chi-squared
estimate to be used in all cases.
correlation
optional argument. If False, the correlation matrix for the coefficients
is not computed (useful for models with many coefficients).
VALUE:
a list is returned with the following components.
call, terms, deviance, null deviance
as contained on object
coefficients
A matrix with three columns, containing the coefficients, their standard errors
and the corresponding t-statistic.
dispersion
the dispersion parameter used in the computations (see
dispersion argument).
df
The number of degrees of freedom for the model and for residuals.
deviance.resid
the deviance residuals, as produced by
residuals(object).
cov.unscaled
The unscaled covariance matrix; i.e, a matrix such that multiplying it
by the dispersion parameter, or estimate thereof, produces an estimated
(asymptotic) covariance matrix for the coefficients.
correlation
The computed correlation matrix for the coefficients in the model.
nas
a logical vector indicating which, if any, coefficients are missing.
DETAILS:
This function is a method for the generic function
for class glm.
It can be invoked by calling
for an object x of the appropriate class, or directly by calling
regardless of the class of the object.
People sometimes call this function to print linear coefficients
and their standard errors for gam models.
This can severly underestimate the standard
errors, because
summary.glm ignores any
colinearity between the nonlinear part of
gam
smooths and other terms.