Previously published experimental data on single-electron stripping for over 100 different neutral-neutral and almost 100 ion-neutral collision pairs are compared with Firsov's statistical ionization model. Reasonable agreement is found for a number of reactions, in particular for collisions between some rare-gas atoms. However, marked disagreement exists for many other collisions. From a model assuming a single-step transition of the ionized electron from the bound state into the continum, a scaling law is developed semiempirically which is found able to reduce a large majority of these published data to within a factor of two of a general cross-section curve. Major disagreements with this model occur only in cases which are particularly well described by the Firsov model.