In constructing the proton distribution functions at low energies, one usually applies Compton-Getting corrections to the observed anisotropies. This implies, among other things, that the spectrum as far as its functional form is concerned has to be defined with respect to a moving frame (solar wind). The justification for applying such corrections, in particular in case of the upstream events, will be discussed. Tools to investigate this 'frame problem' will be described, and some appropriate examples, taken from the many burst events as observed with the low energy proton spectrometer aboard ISEE 3, will be presented. It is found that none of the directional characteristics can be poorly accounted for by Compton-Getting effects.