The factors contributing to the bias introduced in the experimental determination of void fractions in liquid flow systems by methods of radiation diagnosis are being examined. It is shown that error bounds for the radiation transmittance and for the void fraction can be established for certain well-defined void variations. The general validity of the analytical results has been experimentally tested and confirmed. In addition, analytical criteria are derived which relate radiation attenuation and certain parameters of the radiation diagnostic facility to the dynamic voiding effects.