Furthermore, we are skeptical about whether the fair-share interpretation is a
statistical notion at all. It seems to ignore, in a sense, the original distribution of
values and to attend only to the total accumulation of some amount in a group.
Consider, for example, the value we would compute to decide how the different
numbers of candies brought by various children to a party could be equally
redistributed among the children (see Table 1). In this context, the particulars about
how the candies were originally distributed seem irrelevant. That is, the number that
constitutes a fair share is not viewed as a representation or summary of the original
distribution but rather as the answer to the question of how to divide the candies
equitably.