Recent experiments conducted to study the vectorial photoelectric effect with CsI, Al2O3 and Si photocathodes at 2.69 keV indicate null results. Detailed analysis shows that previously measured modulation can be well explained by geometrical misalignment and a combination of the asymmetric shape of the incident X-ray beam and a small detection area of the photoelectron detector. After the elimination of the sources of spurious modulation, we observed a modulation factor of less than 3 percent for a grazing incidence angle as small as 5 deg. There is no observable difference in the pulse height distribution between s and p states.