In solving the regression equations, collinearity in the design matrix can result in inaccurate parameter estimates. The use of orthogonal matrix transformations such as the singular-value decomposition (SVD) can reduce the effect of collinearity. Also, estimates of the regression coefficients can sometimes be improved through iterative refinement. The application of iterative refinement to the SVD solution of the regression equations is described. Results show that iterative refinement using the SVD can improve regression coefficient estimates in the cases where the design matrix is highly collinear.