CRITERIA Subcommand (CSORDINAL command)

The CRITERIA subcommand offers controls on the iterative algorithm used for estimation, and specifies numerical tolerance for checking singularity.

CHKSEP = integer. Starting iteration for checking complete and quasi-complete separation. Specify a non-negative integer. This criterion is not used if the value is 0. The default value is 20.

CILEVEL = value. Confidence interval level for coefficient estimates, exponentiated coefficient estimates, and cumulative odds ratio estimates. Specify a value greater than or equal to 0, and less than 100. The default value is 95.

DF = value. Sampling design degrees of freedom to use in computing p values for all test statistics. Specify a positive number. The default value is the difference between the number of primary sampling units and the number of strata in the first stage of sampling.

LCONVERGE = [number (RELATIVE | ABSOLUTE)]. Log-likelihood function convergence criterion. Convergence is assumed if the relative or absolute change in the log-likelihood function is less than the given value. This criterion is not used if the value is 0. Specify square brackets containing a non-negative number followed optionally by the keyword RELATIVE or ABSOLUTE, which indicates the type of change. The default value is 0; the default type is RELATIVE.

METHOD = FISHER(number) | NEWTON. Model parameters estimation method. The Fisher scoring method is specified by the keyword FISHER, the Newton-Raphson method, by the keyword NEWTON, and a hybrid method is available by specifying FISHER(n). In the hybrid method, n is the maximal number of Fisher scoring iterations before switching to the Newton-Raphson method. If convergence is achieved during the Fisher scoring phase of the hybrid method, iterations continue with the Newton-Raphson method.

MXITER = integer. Maximum number of iterations. Specify a non-negative integer. The default value is 100.

MXSTEP = integer. Maximum step-halving allowed. Specify a positive integer. The default value is 5.

PCONVERGE = [number (RELATIVE | ABSOLUTE)]. Parameter estimates convergence criterion. Convergence is assumed if the relative or absolute change in the parameter estimates is less than the given value. This criterion is not used if the value is 0. Specify square brackets containing a non-negative number followed optionally by the keyword RELATIVE or ABSOLUTE, which indicates the type of change. The default value is 10^-6; the default type is RELATIVE.

SINGULAR = value. Tolerance value used to test for singularity. Specify a positive value. The default value is 10^-12.