The CATMOD Procedure

Overview

The CATMOD procedure now offers the iterative proportional fitting (IPF) algorithm for fitting hierarchical log-linear models with one population (that is, there are no independent variables and no population variables). The advantage of the IPF algorithm is that you can obtain the log likelihood, G², and the predicted cell counts without performing expensive parameter estimation and covariance computations. To request IPF fitting, you specify the ML=IPF option. Several options for controlling the convergence criterion are available, as well as options for controlling how the degrees of freedom for G² are computed.

The new MISSING= option specifies whether a missing cell is treated as a sampling or structural zero. The new ZERO= option specifies whether a non-missing cell with zero weight is treated as a sampling or structural zero.

Chapter Contents
Previous
Next
Top