One-dimensional neutronic calculations in a simple geometrical model, which are used frequently in blanket design and shielding analyses, are qualified by a comparison with three-dimensional calculations in a realistic tokamak model. The Next European Torus (NET) reactor is used as an example of a well-developed design for a “next-step” tokamak machine. Various blanket concepts with different neutronic characteristics are taken into account: a helium-cooled solid breeder blanket with beryllium as neutron multiplier, a self-cooled liquid-metal blanket with the eutectic alloy Pb-17Li, or, alternatively, pure lithium as breeding material/coolant and an aqueous lithium salt solution blanket. The calculations are performed with the MCNP Monte Carlo code, both in the one- and the three-dimensional approach. It is shown that the use of the one-dimensional approach can be justified for design and shielding calculations, if the plasma source is normalized in a consistent manner and both its radial distribution and its angular dependence are chosen appropriately. The latter requirement necessitates the use of an anisotropic neutron source distribution in the one-dimensional calculation. The tritium breeding ratio is overestimated in the one-dimensional approach to a degree that depends on the neutronic characteristics of the blanket variants used. A blanket concept evaluation, therefore, is valid only on the basis of three-dimensional calculations in the actual tokamak geometry. One-dimensional shielding calculations on average agree rather well with three-dimensional ones, although they do not allow “safe” results to be obtained. As the safety margins for the shielding system in general are crucial, a proof by three-dimensional shielding calculations in the real tokamak geometry is required.