The computational data suggests that the “attenuation length” measurement measures something other than the attenuation length of the plastic. Rather, it's a sort of overall attenuation length for the geometry and characterizes the degree to which the unmasked scintillator can collect and detect optical photons. There is a fair amount of agreement between the 2.5m/1.00 simulation and the DMR-15C measurement; even inspecting the points visually yields a very similar muon peak in terms of magnitude. But these simulations were not Gd-loaded, and there is no match between the simulations and the unloaded sample DMR-15A.