The Kellermann study has three major flaws.

1) It included suicides (in effing King County Washington ... the suicide capital of the US).
2) It required you to KILL an intruder in order to count as a "self defensive use".
3) It didn't differentiate between types of owners (a gang banger with a stolen glock is counted equally with an Orthodontist that collects 5 figure Italian shotguns).

So the original finding was that a gun kept in the home is 43 times more likely to kill a family member or acquaintance* than to kill an intruder. (*note that "acquaintance" as used here simply meant that the shooter and shootee knew each other before the shooting ... so a pimp shooting his ho or a drug dealer shooting a delinquent customer is the same thing as an innocent child finding a gun and accidentally shooting one of their siblings).

After Kellermann was called on the carpet for his bogus study even he revised the numbers down to 2.5 to 1 (so I have no idea where this idiot got the 22 to 1 figure). Even at 2.5 to 1 his study is flawed (all he did is removed the suicides ... all the other flaws still exist).