

   GGeenneerraalliizzeedd LLiinneeaarr MMiixxeedd MMooddeellss

        glmm(formula, family=family, data=list(), weights=NULL, offset=NULL,
             nest, delta=1, maxiter=20, points=10, print.level=0,
             control=glm.control(epsilon=0.0001,maxit=10,trace=FALSE))

   AArrgguummeennttss::

    formula: A symbolic description of the model to be fitted.
             If it contains transformations of the data, a
             dataframe must be supplied.

     family: A description of the error distribution and link
             function to be used in the model; see `family' for
             details.

       data: An optional dataframe containing the variables in
             the model.

    weights: An optional weight vector. If this is used, data
             must be supplied in a data.frame.

     offset: The known component in the linear predictor. If
             this is used, data must be supplied in a
             data.frame. An offset cannot be specified in the
             model formula.

       nest: The variable classifying observations by the unit
             (cluster) upon which they were observed.

      delta: If the response variable has been transformed,
             this is the Jacobian of that transformation, so
             that AICs are comparable.

    maxiter: The maximum number of iterations of the outer loop
             for numerical integration.

     points: The number of points for Gauss-Hermite integration
             of the random effect.

   print.level: If set equal to 2, the log probabilities are
             printed out when the underflow error is given.

    control: A list of parameters for controlling the fitting
             process.

   DDeessccrriippttiioonn::

        `glmm' fits a generalized linear mixed model with a
        random intercept using a normal mixing distribution
        computed by Gauss-Hermite integration. For the normal,
        gamma, and inverse Gaussian distributions, the
        deviances supplied are -2 log likelihood, not the usual
        `glm' deviance; the degrees of freedom take into
        account estimation of the dispersion parameter.

        If weights and/or offset are to be used or the formula
        transforms the some variables, all of the data must be
        supplied in a dataframe.  Because the `glm' function is
        such a hack, if this is not done, weird error messages
        will result.

        na.omit is not allowed.

   VVaalluuee::

        `glmm' returns a list of class `glmm'

   AAuutthhoorr((ss))::

        J.K. Lindsey

   SSeeee AAllssoo::

        `family', `glm', `glm.control', `gnlmm', `gnlr',
        `gnlr3', `fmr'

   EExxaammpplleess::

        nest <- gl(5,4)
        y <- rpois(20,5+2*codes(nest))
        # overdispersion model
        glmm(y~1, family=poisson, nest=gl(20,1), points=3)
        # clustered model
        glmm(y~1, family=poisson, nest=nest, points=3)

