The flow ansatz states that the single-particle distribution of a given event can be described in terms of the complex flow coefficients V_n . Multi-particle distributions can therefore be expressed as products of these single-particle coefficients; a property commonly referred to as factorization. The amplitudes and phases of the coefficients fluctuate from event to event, possibly breaking the factorization assumption for event-sample averaged multi-particle distributions. Furthermore, non-flow effects such as di-jets may also break the factorization assumption. The factorization breaking with respect to pseudorapidity η provides insights into the fluctuations of the initial conditions of heavy ion collisions and can simultaneously be used to identify regions of the phase space which exhibit non-flow effects. These proceedings present a method to perform a factorization of the two-particle Fourier coefficients V _nΔ(η_a , η_b ) which is largely independent of detector effects. AMPT model calculations of Pb-Pb collisions at √sNN = 5.02 TeV are used to identify the smallest |Δη|-gap necessary for the factorization assumption to hold. Furthermore, a possible Δη-dependent decorrelation effect in the simulated data is quantified using the empirical parameter . The decorrelation effect observed in the AMPT calculations is compared to results by the CMS collaboration for Pb-Pb collisions at √sNN = 2.76 TeV.