Calculation of the neutron flux in a nuclear reactor core is ideally performed by solving the neutron transport equation for a detailed-geometry model using several tens of energy groups. However, performing such detailed calculations for an entire core is prohibitively expensive from a computational perspective. Full-core neutronic calculations for CANDU reactors are therefore performed customarily using two-energy-group diffusion theory (no angular dependence) for a node-homogenized reactor model. The work presented here is concerned with reducing the loss in accuracy entailed when going from Transport to Diffusion. To this end a new method of calculating the diffusion coefficient was developed, based on equating the neutron balance equation expressed by the transport equation with the neutron balance equation expressed by the diffusion equation. The technique is tested on a simple twelve-node model and is shown to produce transport-like accuracy without the associated computational effort. / UOIT