Using factorization we determine the allowed domains of the ratios of form factors, x = A 2 (0)/A 1 (0) and y = V (0)/A 1 (0), from the experimentally measured ratio R h ≡ Γ(D + s → φρ + )/Γ(D + s → φπ + ) assuming three different scenarios for the q 2 -dependence of the form factors. We find that the allowed domains overlap with those obtained by using the experimentally measured ratio R sℓ = Γ(D + s → φℓ + ν ℓ )/Γ(D + s → φπ + ) provided that the phenomenological parameter a 1 is 1.23. Such a comparison presents a genuine test of factorization. We calculate the longitudinal polarization fraction,, in the three scenarios for the q 2dependence of the form factors and emphasize the importance of measuring Γ L /Γ. Finally we discuss the q 2 -distribution of the semileptonic decay and find that it is rather insensitive to the scenarios for the q 2 -dependence of the form factors, and unless very accurate data can be obtained it is unlikely to discriminate between the different scenarios. Useful information on the value of x might be obtained by the magnitude of the q 2 -distribution near q 2 = 0. However the most precise information on x and y would come from the knowledge of the longitudinal and left-right transverse polarizations of the final vector mesons in hadronic and/or semileptonic decays.