Positron annihilation lifetime spectra of amorphous polymers are commonly deconvoluted into three lifetime components (3C analysis). For high quality lifetime spectra, a deconvolution into four discrete or distributed lifetime components (4C analysis) is sometimes required to achieve an acceptable variance of the fit. Dlubek et al, [1] have pointed out the possibility that the two longest lifetime components of a 4C analysis result from a splitting up of one broad ortho-positronium (o-Ps) lifetime distribution into two narrow sub-peaks in the ns-lifetime range. A second interpretation can be obtained out of the analysis of simulated lifetime spectra, which contain information on the slow trapping (ST) process of Ps. This process is characterized by a slow relaxation of the time dependent Ps decay rate lambda (t); from lambda (0), the decay rate at the time of Ps formation in the occupied volume of the polymer, to lambda(infinity), the decay rate of the fully trapped Ps into the free volume of the polymer. Since the simulated spectra can be fitted perfectly into four discrete or distributed lifetime components, the two longest lifetime components of a 4C analysis can also be due to a false interpretation of this ST process. However, positron lifetime experiments even with very high statistics (10(8) counts) can make no distinction between a broad o-Ps lifetime distribution and the slow trapping process of Ps.