“…Part of the modeling done up to now estimates global trends related to radioactivity samples taken in inhabited areas (Durrani et al, 1997;Lac an et al, 2006;Rybach et al, 2002). The usual way to model indoor radon concentrations is to assume normal or more often lognormal distributions of concentrations on a given territory (Miles and Ball, 1996;Miles, 1998;Zhu et al, 2001), where eventual extreme values are omitted from the data set or used with statistical procedures that give the extremes a limited importance.…”