starkshift

Does someone have a good model for estimating these kinds of figures given widely divergent reports in press sources and otherwise? How might we approach such a problem with a greater degree of statistical/mathematical…