The chain scission rate constant and the initiator decomposition efficiency are very important kinetic parameters in the chemically initiated degradation of polypropylene. These two kinetic parameters have been estimated using both deterministic and stochastic modelling procedures. To relate the chain scission rate constant to measurable properties of this system in a deterministic model, the stationary state hypothesis for the peroxide radical concentration has been relaxed and a new model relating both the overall chain scission rate constant and the peroxide decomposition efficiency to three average molecular weights has been formulated. Using Monte Carlo simulation results, the sensitivity of the kinetic parameter posterior probability surface to the reaction time has been investigated. The results indicate that there is an optimum time for collection of experimental data in the reaction system so that the estimated values of the kinetic parameters are more reliable. Based on simulated data at a reaction time of 30 seconds, the point estimates and the joint confidence region have been obtained for the chain scission rate constant (k(1)) and the peroxide decomposition efficiency (f) using a Bayesian approach to statistical inference.