Judging by the ranges, it's my guess that these are not extremes, but possibly 25/75 quartile marks. The methodology isn't clear, so we don't know how the data

Judging by the ranges, it's my guess that these are not extremes, but possibly 25/75 quartile marks. The methodology isn't clear, so we don't know how the data were aggregated/culled, but it seems that +/- 3 SD data, at least, were stripped in creating these tables.