In traditional Hall effect experiments for Hall sensitivity measurement, 4 side effects are usually considered. Based on the experimental data, the influence of side effects on the accuracy of Hall sensitivity is valuated. It is confirmed that the voltage drop due to unequal potential is most prominent, while other side effects are negligible. Furthermore, considering that the Hall voltage and the voltage drop due to unequal potential vary differently with the magnitudes of the working current and the external magnetic field, two new schemes, i.e., applying linear least squares method and binary quadratic nonlinear model, respectively, for measuring Hall sensitivity are proposed. Then, the strategy for eliminating system errors with symmetric measurement method in the traditional experiment is improved to a strategy of solving problems from differences.