The coupling of physical separation and chemical reactions in one unit operation, the so-called reactive separation processes ({RSPs}), are of increasing interest for scientific investigation and industrial application. {RSPs} tend to be very complex due to strong physico-chemical interactions, and a number of the model assumptions that are adequate for traditional unit operations may not be suitable in this new context. Models used to simulate RSPs should therefore be highly accurate in order to reflect this complexity. On the other hand, it is not always economical nor feasible to use the most sophisticated models for a range of different applications, such as process and equipment design, the determination of optimal operating policies, model-based control, etc., and the model complexity has to be adapted to the purpose. In many cases however, it is not evident which simplifications are appropriate, so that the uncertainty concerning adequate ways to model {RSPs} is still significant. Kreul et al. (1996a, b) have published a very detailed rate-based model and first simulation results for RSPs including reaction kinetic terms in the bulk and the film areas. In this paper, a detailed model investigation is presented, as well as a systematic deduction of possible model simplifications. The completely kinetic model and the corresponding equilibrium stage model are applied to various homogeneously catalyzed reactive distillation and reactive absorption processes. A number of simulation results are presented for four very different test systems. It is concluded that equilibrium and rate-based approaches can lead to significantly larger differences in calculated concentrations profiles for RSPs than for non-reactive operations, so that the additional effort of more complex modeling may be justified. In addition, it is demonstrated that if the kinetics of the mass and energy transfer between the phases are calculated explicitly, in most cases the consideration of interaction phenomena such as diffusional and direct reaction-transfer interaction is not necessary.