Thursday, January 15, 2009

Attack an insider, kill an outsider? I've read that animals as a rule exhibit non-lethal violence toward members of the in-group and lethal violence toward outsiders, and I wondered if it might apply to human races.

I found rates of within- and cross-race aggravated assaults and homicides for 2006 and looked to see if assaults turn into murders more frequently if the attacks are cross-race. Here are the ratios of aggravated assaults to homicides for different combinations:

Ratio of aggravated assaults to homicides

White-on-white 172.0
Black-on-black 55.1
White-on-black 148.5
Black-on-white 108.7

Just to make sure you're interpreting those numbers right, using the first one, read it as, "For whites attacking whites, you'll have 172 aggravated assaults for every murder." So the lower the number, the more likely someone ends up dead.

So, you can see that black-on-black violence is the deadliest; white-on-white the least so. The second deadliest is black on white, so the chance of death is highest when the perpetrator is black.

Whites are somewhat more likely to kill a black victim than a white one, so this is consistent with the hypothesis, but a black attacker is much more likely to finish off the victim if he is also black--a non-supportive finding.

I should mention the important fact that a great deal of serious violence is mutual combat. The story here seems to be that black participants in violence--whether they end up the victim or the aggressor--are more intensely violent than whites. Have two whites go at it, and you get few deaths. Insert a black as the victim or especially the offender, you get more deaths. Insert two blacks, and the chance of murder jumps even higher.

Does this seem right, or are there other ideas?


  1. Anonymous7:30 AM

    Weapons such as guns and knives frequently cause assaults to become homicides. If certain racial pairings are more likely than others to involve weapons, it's not at all surprising that the former type have a higher death rate. I don't have the time right now to look at the original source documents, do they account for the presence of weapons?


  2. No, but you're right that guns may play a role here. Roughly 70% of homicides involve a gun, but this number is even higher for gang-related murders--often black on black.

    On the other hand, the use of a gun might simply reflect greater lethal intent.

  3. Anonymous9:48 AM

    "On the other hand, the use of a gun might simply reflect greater lethal intent."

    You bet it does. Years ago, many times a criminal wouldn't even have the gun loaded. Now, it is go in shooting. You'll get popped first, then robbed as you bleed out.

  4. Benji8:37 PM

    White and hispanic categories are conflated in crime statistics. The numbers will not be reliable unless you can separate the categories.

    The ratio you present means very little without the raw data. With the raw data you can calculate the number of deaths in each category.

    The concept of "outsider" means much more than color of skin. It could mean someone from another gang or another neighborhood.

    Blood feuds are more likely to start between members of the same race who have interactions that go badly. A person can very quickly be categorized as "other".

  5. A lot of black on white assaults are probably robberies. The assailant is likely not to kill his victim unless there is resistance.

    A higher percentage of black on black is probably gang wars or disrespect. Death is a natural outcome there.


Which oil for cooking is the healthiest? A meta-analysis of 54 trials

This new meta-analysis of 54 randomized studies looked to see which oils improve your cholesterol the best. The authors found that you ca...