Cold plasma theory and parallel wave propagation are often assumed when approximating the whistler mode magnetic field wave power from electric field observations. The current study is the first to include the wave normal angle from the Electric and Magnetic Field Instrument Suite and Integrated Science package on board the Van Allen Probes in the conversion factor, thus allowing for the accuracy of these assumptions to be quantified. Results indicate that removing the assumption of parallel propagation does not significantly affect calculated plasmaspheric hiss wave powers. Hence, the assumption of parallel propagation is valid. For chorus waves, inclusion of the wave normal angle in the conversion factor leads to significant alterations in the distribution of wave power ratios (observed/ calculated); the percentage of overestimates decreases, the percentage of underestimates increases, and the spread of values is significantly reduced. Calculated plasmaspheric hiss wave powers are, on average, a good estimate of those observed, whereas calculated chorus wave powers are persistently and systematically underestimated. Investigation of wave power ratios (observed/calculated), as a function of frequency and plasma density, reveals a structure consistent with signal attenuation via the formation of a plasma sheath around the Electric Field and Waves spherical double probes instrument. A simple, density-dependent model is developed in order to quantify this effect of variable impedance between the electric field antenna and the plasma interface. This sheath impedance model is then demonstrated to be successful in significantly improving agreement between calculated and observed power spectra and wave powers.