Tim Lambert (deltoidblog AT gmail.com) is a computer scientist at the University of New South Wales.
Method of % Completed % Attacked % Injured Num Times Self Protection Used(a) Used gun 30.9 25.2 17.4 89,009 Used Knife 35.2 55.6 40.3 59,813 Used other weapon 28.9 41.5 22.0 104,700 Used physical force 50.1 75.6 50.8 1,653,880 Tried to get help or frighten offender 63.9 73.5 48.9 1,516,141 Threatened or reasoned with offender 53.7 48.1 30.7 955,398 Nonviolent resistance, including evasion 50.8 54.7 34.9 1,539,895 Other measures 48.5 47.3 26.5 284,423 Any self-protection 52.1 60.8 38.2 4,603,671 No self-protection 88.5 41.5 24.7 2,686,960 Total 65.4 53.7 33.2 7,290,631
T. Mark Gibson said:
The above table clearly shows that one may best improve ones odds of surviving a criminal attack uninjured by defending one's self with a gun.
This table shows that "Used gun" is correlated with lower injury rates. If two factors A and B are correlated, then it is possible that A causes B, B causes A, some third factor causes A and B, or any combination.
In this example, we can find plausible explanations involving all three possibilities.
A causes B: Using a gun for self-protection lowers your chance of injury in a robbery.
B causes A: Injury makes it harder for you to use a gun for self-protection.
Third factor causes A and B: Preparing for possible robberies (for example by being alert to your surroundings) lowers your chance of injury. Part of that preparation may include carrying a gun.
It may be comforting to believe that carrying a gun can reduce your chance of injury in robbery, but it is best not to let wishful thinking cloud your judgment on this issue.
By the way, the data in the above table comes from the National Crime Survey. I'm surprised that all the posters who loudly proclaimed that the NCS was worthless for measuring gun use for self protection have fallen silent. Why aren't they telling us that this table is worthless for measuring the effectiveness of gun use for self protection?
