Reliable experimental determination of the mean effective charge (Zeff) of the plasma is of enormous importance for the impurity control in high temperature plasmas, allowing the performance of various discharge scenarios to be assessed and the benchmarking of model calculations, ideally accompanied by measurements of the Zeff profile. The Zeff is usually calculated using data from line-integrated visible bremsstrahlung measurements, electron density and temperature profiles, and the plasma geometry. On Tore Supra (TS) it had been noticed in the past that hot spots on the inner wall, MARFEs or supra thermal electrons can falsify the signal of the visible bremsstrahlung. More significant, however, was a systematic overestimation of the edge channels, giving unreasonably high edge Zeff values of 30 or even higher. The analysis of dedicated discharges in the present configuration of TS leads us to conclude that this is due to the influence of vessel wall reflections. A newly developed inversion method which takes reflections into account shows that the presence of reflections leads to an overestimation of the outer channels in the reconstruction of the Zeff profile. This is due to the larger relative contribution of the reflections on the edge channels due to the shorter path length through the plasma and the lower bremsstrahlung emission at the edge. We present a model which takes into account this effect and allows a correction procedure to be applied when calculating the Zeff profile. The TS data can be modeled assuming Lambertian (diffusive) reflections with a reflectivity coefficient of 20%-25%, which is in agreement with measurements of the reflectivity of the inner wall of the TS vacuum chamber.