We explored the transition of 13 X-linked markers across two separate portions of the house mouse hybrid zone, asking whether such a comparison can distinguish the effects of selection from random factors. A heuristic search in the likelihood landscape revealed more complex likelihood profiles for data sampled in two-dimensional (2D) space relative to data sampled along a linear transect. Randomized resampling of localities analyzed for individual loci showed that deletion of sites away from the zone center can decrease cline width estimates whereas deletion of sites close to the center can significantly increase the width estimates. Deleting localities for all loci resulted in wider clines if the number of samples from the center was limited. The results suggest that, given the great variation in width estimates resulting from inclusion/exclusion of sampling sites, the geographic sampling design is important in hybrid zone studies and that our inferences should take into account measures of uncertainty such as support intervals. The comparison of the two transects indicates cline widths are narrower for loci in the central part of the X chromosome, suggesting selection is stronger in this region and genetic incompatibilities may have at least partly common architecture in the house mouse hybrid zone.