

   SSpplliitt--pplloott AANNOOVVAA MMooddeell

        rmaov(response, tvcov=NULL, ccov=NULL, analysis=T)

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

   response: A matrix or dataframe of response values with
             units as rows and repeated measures as columns.

      tvcov: A numeric vector or factor variable defining the
             clusters. If there are several levels of nesting,
             a matrix or dataframe with columns of such vari-
             ables defining the nested clusters starting with
             the highest level (that is, from slowest to
             fastest varying). If not provided, each response
             value of a unit is assumed to belong to a differ-
             ent cluster (that is, one factor with
             `ncol(response)' levels is assumed).

       ccov: A vector or factor variable for one inter-subject
             covariate or a matrix, dataframe, or list of sev-
             eral such variables.

   analysis: If FALSE, the design matrix is set up, but the
             analysis is not performed.

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

        `rmaov' performs the classical balanced split-plot
        ANOVA, with `summary' providing the table. This is the
        so-called repeated measures ANOVA.

        For unbalanced data, `elliptic' will perform the analy-
        sis for one or two levels of nesting.

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

        The fitted model is returned.

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

        Ralf Goertz (ralf.goertz@uni-jena.de)

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

        `carma', `elliptic', `lm', `potthoff'.

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

        # vision data for 7 individuals, with response a 7x8 matrix
        # two levels of nesting: 4 levels of power for each eye
        y <- matrix(rnorm(56),ncol=8)
        tvc <- data.frame(eye=c(rep(1,4),rep(2,4)),power=c(1:4,1:4))
        summary(rmaov(y, tvc))

