Differences between lattice Monte Carlo data and perturbation theory (for example the lack of asymptotic scaling) are usually associated with the 'bad' behaviour of the bare lattice coupling g0 due to the effects of large (and unknown) higher order terms in g0. In this philosophy a new, renormalised coupling g ′ is defined with the aim of making the higher order coefficients of the perturbative series in g ′ as small as possible.In this paper an alternative scenario is discussed where lattice artifacts are proposed as the cause of the disagreement between Monte Carlo data and the g0-perturbative series. We find that with the addition of a lattice artifact term, the usual asymptotic scaling expression in g0 is in excellent agreement with Monte Carlo data. Lattice data studied includes the string tension, the hadronic scale r0, the discrete beta function, Mρ, fπ and the 1P-1S splitting in charmonium.