Better dispersion numbers in the tables

[nikohansen]

The dispersion number (see e.g. here) is currently a single number, which is the 10th to 90th interpercentile range.

It would be (much) nicer to have it in two numbers, the 10%tile minus the mean and the 90%tile minus the mean. These could be given on top of each other, that is, instead of $493 (797)$ we see $493 ^{+597}_{-200}$.