We analyze the high-electric-field technique designed by Yokoyama and van Sprang [J. Appl. Phys. 57, 4520 (1985)] to determine the polar anchoring coefficient W of a nematic liquid crystal-solid substrate. The technique implies simultaneous measurement of the optical phase retardation and capacitance as functions of the applied voltage well above the threshold of the Frederiks transition. We develop a generalized model that allows for the determination of W for tilted director orientation. Furthermore, the model results in a new high-field technique, (referred to as the RV technique), based on the measurement of retardation versus applied voltage. W is determined from a simple linear fit over a well-specified voltage window. No capacitancemeasurements are needed to determine W when the dielectric constants of the liquid crystal are known. We analyze the validity of the Yokoyama–van Sprang (YvS) and RV techniques and show that experimental data in real cells often do not follow the theoretical curves. The reason is that the director distribution is inhomogeneous in the plane of the bounding plates, while the theory assumes that the director is not distorted in this plane. This discrepancy can greatly modify the fitted value of 1/W, and even change its sign, thus making the determination of W meaningless. We suggest a protocol that allows one to check if the cell can be used to measureW by the YvS or RV techniques. The protocol establishes new criteria that were absent in the original YvS procedure. The results are compared with other data on W, obtained by a threshold-field technique for the same nematic-substrate pair.