PROC ORTHOREG performs linear least squares regression by using the Gentleman-Givens computational method (Gentleman, 1972, 1973), and it can produce more accurate parameter estimates for ill-conditioned data. PROC GLM and PROC REG produce very accurate estimates for most problems. However, if you have very ill-conditioned data, consider using PROC ORTHOREG. The collinearity diagnostics in PROC REG can help you determine whether PROC ORTHOREG would be useful.