The use of polarimeter in nuclear fusion is increasing so much that ITER will mount three polarimeters for plasma control and diagnostics, while the Divertor Tokamak Test will have a poloidal polarimeter with more than ten lines of sight. Several studies have been conducted on this diagnostic, where the main experimental test bench was the FIR interferometer-polarimeter at Joint European Torus. Polarimetry measurements can be used for different purposes, such as real-time control of the plasma, machine protection, and plasma equilibrium constraint. However, even if the rate of change of the polarisation is strongly related to plasma characteristics, the exact equations that link the plasma quantities (electron density and magnetic field) with the beam polarisation are not linear. Thus, extrapolation of the plasma quantities requires an inversion that is not possible a priori since it would need the knowledge of the measuring quantities. For these reasons, polarimeters should be designed to work under the so-called type-I approximation, which ensures linearity between the line-integrated plasma properties and the polarisation state of the electromagnetic wave. However, the range of validity of the type-I approximation, chosen the laser wavelength, is limited to a specific range of plasma quantities. In this work, the authors proposed a new approach to calculate those line-integrated quantities for poloidal polarimeters. The new approach is developed for two situations, back-reflected and not back-reflected beams. The two proposed methods will be introduced analytically and tested numerically, showing that they can provide more accurate measurements for a wider range of plasma operations.