The dissociative adsorption of oxygen on the stepped Pt(533) surface has been investigated using supersonic molecular beams by measuring the initial dissociative sticking probability S-0 as a function of incident kinetic energy E-i, angle Phi, surface temperature T-S, and S as a function of coverage Theta. By comparison with dynamical data available on the Pt(111) surface we have been able to establish that step sites dominate the dissociative adsorption process. S-0(E-i) for oxygen on Pt(533) at T-S=200 in the energy range 52 meV-1.4 eV shows a similar functional dependence to results on Pt(111), however, the magnitude of S-0 on Pt(533) is significantly greater at all energies. The measurement of S-0(Phi), scattering in a plane perpendicular to the step direction at E-i=1.18 eV at T-S=350 K, reveals a strong and asymmetric angular dependence which contains contributions associated with activated adsorption and dissociation of the chemisorbed precursor on the (111) terraces, and a second contribution associated with activated dissociation through a similar channel at the step sites. The latter exhibits a maximum in S-0(Phi) at 35 degrees, near the angle corresponding to the normal of the (100) step plane. S-0(T-S) at E-i=1.18 eV and Phi =0 degrees reveals a much smaller temperature dependence in the range 150 >T-S(K)> 800 on Pt(533) than on Pt(111). At E-i=1.18 eV and Phi =0 degrees ca. 15%-25% of dissociation takes place through molecules impinging directly at step sites. The remaining fraction dissociate through activated adsorption of the chemisorbed species on the (111) terrace and subsequent partition between desorption, and dissociation at step sites. Dissociation of the chemisorbed precursor on the (111) terrace appears highly activated, a result which is consistent with theory. The rapid decrease in S-0(E-i) observed below 0.15 eV on Pt(533), observed also on Pt(111), is consistent with a trapping mechanism where the need to dissipate energy limits the probability of adsorption, and subsequent dissociation, of the physisorbed precursor. Kinetic modelling of this partition on Pt(533), between the conversion of the physisorbed precursor to the chemisorbed species, and desorption yields DeltaE=120 meV and v(d)/v(pc)=80. We conclude that the effective barrier to conversion of the physisorbed to chemisorbed species on Pt(533) is effectively zero. We conclude that defects will tend to dominate this conversion process on the close packed surface. In addition to this channel, at E-i=0.05 meV ca. 50% of molecules dissociate through the same channel operating at higher energy on Pt(533).