Measurements of quantum corrections to the magnetoconductivity in two superlattice samples of GaAs/${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{\ensuremath{-}}\mathit{x}}$As differing in the ${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{\ensuremath{-}}\mathit{x}}$As alloy composition (and therefore in the miniband width) are reported. The sample with a smaller bandwidth displayed a larger correction to the magnetoconductivity; however, this result cannot be described by a two-dimensional theory. Using both a systematic and a detailed analysis of the transport data, we find that weak-localization (WL) rather than electron-electron interactions dominates the quantum transport corrections. An extensive quantitative interpretation of the transport results is done with an advanced weak-localization theory for superlattices, incorporating dopant distribution, wave-function modulation, and a higher magnetic-field range beyond the eikonal approximation. However, this advanced theory still assumes a \ensuremath{\delta}-function impurity scattering potential. The theoretical fits to the data are not compatible with the as-grown impurity profiles, but are best fit with an effective uniform impurity distribution. While silicon dopant migration smears the impurity profile, we think some effect comes from the inadequacy of the point-scattering assumption commonly adopted in WL theory, which suggests that the long-range potential scattering in WL effects is operative.