In standard Mott electron polarimetry using thin gold film targets, extrapolation procedures must be used to reduce the experimentally measured asymmetriesAto the values they would have for scattering from single atoms. These extrapolations involve the dependence ofAon either the gold film thicknesstor the maximum detected electron energy loss in the target &Dgr;E. Using a concentric cylindrical‐electrode Mott polarimeter, we have studied and compared these two types of extrapolations over the electron energy range 20–100 keV. The potential systematic errors which can result from such procedures are analyzed in detail, particularly with regard to the use of various fitting functions in thickness extrapolations, and the failure of perfect energy‐loss discrimination to yield accurate polarizations when thick foils are used. A critical discussion of previous work on this subject is presented.