Stratospheric mass spectrometry of positive ions using a passive chemical ionization method is biased by collisionally induced dissociation (CID) of the sampled cluster ions behind the inlet hole of the spectrometer. This mechanism is investigated by means of a one-dimensional model for clusters of the family H+ (CH3CN)l(H2O)m which are involved in the volume mixing ratio determination of acetonitrile in the stratosphere. The three physical parameters upon which CID depends are outlined: the density of the neutral gas jet, the velocity or barycentric kinetic energy raised by the ions in the static electric field and the total cross-section for the dissociation process. The CID effect is modelled by using related simple neutral gas adiabatic expansion laws, a mobility concept for the ions and various experimental data for the CID cross-sections. A macroscopic calculation is completed with no allowance for ion velocity distribution, and a more realistic Monte Carlo computation is also performed. The experimental acetonitrile concentration is found to be underestimated by a typical factor of four at an altitude of 25 km. The sensitivity of our results to the lack of experimental data is explored and some criticism is developed of the confidence limits of such a model.