Multidimensional linear regression analysis is employed as a modeling technique for the prediction of boiling water reactor (BWR) shutdown margin reactivity. A comparison is made between the best models developed using regression analysis and the General Electric (GE) three-dimensional BWR core simulator code. The GE code is based on one-group diffusion theory, and its accuracy is verified by comparison with experimental data. One use of this code is the calculation of shutdown margin throughout a fuel cycle, but it requires a large computing facility not located at a reactor site. The regression models give an approximation of a core's shutdown margin based on current core physics parameters. The method can be utilized at a BWR plant site to provide information demonstrating compliance with license and technical specification requirements. The results obtained by regression predictions for the two cores studied compare favorably with current industry methods. After establishing a regression model, predictions can be made at a reactor site using a pocket calculator.