Appearance potential mass spectrometry (APMS) has recently gained importance for detection and quantitative measurements of reactive radical species in plasmas using line-of-sight sampling of radicals. In this work, we have investigated the assumption that the extraction efficiency of ions produced by direct ionization of radicals, and ions produced by dissociative ionization of the parent molecule used as the reference signal, are equal in the ionizer of the mass spectrometer. We find that the dissociative ionization products are extracted with much lower efficiency (2-50 times smaller for the cases studied) than the direct ionization products. This is expected due to the excess kinetic energy of the dissociatively ionized products as a result of the Franck-Condon effect. Use of this procedure will thus lead to an overestimation of the radical number density by a factor of 2-50, depending on the nature of the parent and the daughter ion. We recommend an alternate procedure for APMS that utilizes an inert gas direct ionization signal as a reference for calibration of the radical signal to calculate the radical number density. The biggest uncertainty in the radical number density estimation with the recommended procedure is expected to be the uncertainty in the published cross sections.