“…The long-term psychological and physical trauma caused by female genital mutilation (FGM) and the world-wide advocacy to eradicate the practice have led researchers to adopt different analytical procedures to examine the origin, types and factors that favour the practice, especially in the developing world (Foster, 1992;Obi, 2004;Jeremiah et al, 2014;Setegn et al, 2016). However, since FGM data are usually collected over several locations and thus geo-referenced, most of the adopted methods, especially classical linear regression or correlation estimations, assume that the relationships being modelled are the same everywhere, thereby estimating, for each variable under consideration, a single value for all study locations.…”