A polytropic model is used to investigate the effects of dark photospheric spots on the evolution and radii of magnetically active, low-mass (M<0.5Msun), pre-main sequence (PMS) stars. Spots slow the contraction along Hayashi tracks and inflate the radii of PMS stars by a factor of (1-beta)^{-N} compared to unspotted stars of the same luminosity, where beta is the equivalent covering fraction of dark starspots and N \simeq 0.45+/-0.05. This is a much stronger inflation than predicted by the models of Spruit & Weiss (1986) for main sequence stars with the same beta, where N \sim 0.2 to 0.3. These models have been compared to radii determined for very magnetically active K- and M-dwarfs in the young Pleiades and NGC 2516 clusters, and the radii of tidally-locked, low-mass eclipsing binary components. The binary components and ZAMS K-dwarfs have radii inflated by \sim 10 per cent compared to an empirical radius-luminosity relation that is defined by magnetically inactive field dwarfs with interferometrically measured radii; low-mass M-type PMS stars, that are still on their Hayashi tracks, are inflated by up to \sim 40 per cent. If this were attributable to starspots alone, we estimate that an effective spot coverage of 0.35 < beta < 0.51 is required. Alternatively, global inhibition of convective flux transport by dynamo-generated fields may play a role. However, we find greater consistency with the starspot models when comparing the loci of active young stars and inactive field stars in colour-magnitude diagrams, particularly for the highly inflated PMS stars, where the large, uniform temperature reduction required in globally inhibited convection models would cause the stars to be much redder than observed.