Misleading Statistics

This story from Slate Magazine illustrates one of the many, many, many ways statistics can mislead:

When the categories get conflated, the statistics can become confusing. Take the number 800,000: It's true that 797,500 people under 18 were reported missing in a one-year period, according to a 2002 study. But of those cases, 203,900 were family abductions, 58,200 were nonfamily abductions, and only 115 were "stereotypical kidnappings," defined in one study as "a nonfamily abduction perpetrated by a slight acquaintance or stranger in which a child is detained overnight, transported at least 50 miles, held for ransom or abducted with the intent to keep the child permanently, or killed." Even these categories can be misleading: Overstaying a visit with a noncustodial parent, for example, could qualify as a family abduction.
Anytime you see such a high number, it pays to remember that 50,000 is the magic number in public policy discussions (example 1, example 2).

Public radio show On the Media puts it best:

Numbers justify fear. 50,000 abducted children, for example, or 50,000 predators prowling for children online. That last figure appeared in a recent introduction for NBC’s “Dateline.� And last week, Attorney General Alberto Gonzales cited Dateline’s number. But where did it come from? So far as statistics go, it turns out that 50,000 is something of a Goldilocks number in the media – not too big and not too small, but for scaring the public - just right.