GENREG <data>,L
estimates a generalized linear model (GLM) defined by ERROR and LINK speci-
fications from a Survo data where the dependent variate is acti-
vated by Y and the covariates by X's. IND and CASES specifications
are used for selection of observations.
The theory of GLM's is described in 'Generalized Linear Models' by
P.McCullagh and J.A.Nelder (Chapman and Hill 1983).
Variables for the model and for results are selected by masks as follows:
Y dependent variate (only one Y allowed)
X covariate (also the constant term must be specified by X)
R for output of scaled Pearson residuals (optional)
P for predicted values of Y (optional)
N number of trials in ERROR=Binomial (Y is number of successes)
E = ERROR specification
L = LINK specification
P = Printout of results
T = Using TABFIT as a preparatory operation for GENREG
S = General information on statistical operations
Help System (web edition)