SECTION 1.8 T-RATIOS

This routine computes the ratio of each coefficient estimate to the square root of its asymptotic variance. If NOBS (see Section 1.4) equals zero, the problem is assumed to be a maximum likelihood problem. If it is positive, a sum-of-squares minimization problem is assumed. (Zero is the default value). To obtain T-Ratios, include these statements in your main program:

COMMON/BTRAT/ITRFLG

ITRFLG=1

Note: T-Ratios are available only with GRADX and DFP. T-Ratios require an extra NP*NP words of storage in the stack. Setting ITRFLG = 0 turns T-Ratios off.