They count the death from firearms taking place in the owner's home, which is a real specific thing not including prevented muggings for instance. And they include not just accidental shootings but suicides and murder of family members, well you're always way more likely to kill or be killed someone you know well, and it seems damned dishonest to imply keeping a gun in your house for robbers is more likely to result in little Timmy getting Excedrin Migraine Headache #357 from playing with the gun or getting shot by a spooked dad then a robber. Not when most of those 'household deaths' will be suicide and most of the remainder will be 'expedited divorces'.

Kellerman's BS '43 times more likely' study has been debunked so many times by so many groups it's absurd. HE never included cases where the trigger was not pulled and the intruder simply detained till cops arrived, he repeatedly refused to allow others to check his data, he included in home by gun suicide, and it goes on. Yet people here the narrative so often it sinks in.