I just read an article on whether people in their homes should be free to use lethal force against burglars even when they don't have specific reason to fear death, serious bodily injury, rape, or kidnapping from the burglar. One question that arises, of course, is whether they have enough general reason to fear injury from such crimes -- just on the theory that someone who will break into their home may also attack them further -- that they should be allowed to assume that the burglar may mean them harm, and use lethal force on that assumption. Here's the relevant discussion:

[B]urglary, from a statistical standpoint, does not appear to be a particularly violent crime. The vast majority of burglars appear to be unarmed, and the vast majority of burglaries involve actual or attempted theft of household property, rather than violent crimes.... Indeed, the incidence of violent attacks by intruders during burglary is quite small. Violent crimes (generally defined as rape, robbery, and assault) were committed during only about 3.8% of all burglaries committed during a ten-year period studied by the Department of Justices.

Sounds like a pretty telling statistic, no? But then let's look at the small print in the footnote:

It should be noted, however, that when a household member was present, the incidence of violent crimes was considerably higher -- about 30.2%.

It's to the article's credit that it mentions this fact, but shouldn't this go in the text, instead of the 3.8% statistic, rather than in the footnote? After all, the percentage gotten when the denominator is only those crimes where there's someone whom the burglar to attack (30.2%) is rather more relevant than the percentage gotten when the denominator is all burglaries.