A radio polarization study of gamma-ray-detected pulsars reveals a surprising tendency for the magnetic and rotation axes to be relatively aligned. This provides tension with gamma-ray models, which disfavour such alignment. The lack of correlation between these findings and those derived from the gamma-ray light curves suggests problems in the models. To make the data consistent with a random orientation of the magnetic field the emission regions could be assumed to extend outside what is traditionally thought to be the open-field-line region in a magnetic inclination angle dependent way. Both acceptance and rejection of this hypothesis has important consequences. Finally, a unification scheme is proposed to explain the observational differences between gamma-ray loud and gamma-ray quiet radio pulsars. This unification scheme takes the orientation of the line of sight and the magnetic inclination angle to be key parameters affecting both the radio and gamma-ray light-curve morphology.