Hanson breaks new ground—NOT; and, the question of interracial violence rates

Aditya B. writes:

In an unbelievable example of honesty and courage, Victor Davis Hanson has, in Steyn-like fashion, tentatively inquired whether racial stereotyping by the police is in fact justified. He even states, without any equivocation, that the majority is more likely to be a victim of the minority than vice versa.

LA replies:

I think you are overstating the importance and courage of Hanson’s column. He is not breaking any new ground here that I can see. When he says that that the majority is more likely to be a victim of the minority than vice versa, he is not saying that blacks are more violent than whites, he is making an arithmetical statement that would be true even if the majority and minority had equal rates of violence. If group A is six times larger than group B, and if both groups rates of intergroup violence is the same, then there would be six times the number of B-on-A violence than A-on-B violence.

In reality the black-on-white violence rate is not merely six times the white-on-black violence rate, but more like 50 times. The reason for this is that in addition to whites being a six-times larger group than blacks, which automatically makes black-on-white violence six times greater than white-on-black violence, blacks are between eight and ten times more violent than whites, so the result is six times eight or about 50 times greater.

Aditya replies:

So basically, it was another weasel piece. Oy! These people will never learn.

March 27

Several readers have told me that my above remark on the data of interracial violence is incorrect. When I wrote it, I was roughly recalling an issue I have not thought about in detail in years. Before I can post the readers’ comments and reply adequately, I will have to focus on the mathematical questions, which I have not had time to do what with the raging Martin-Zimmerman issue and others being addressed at the site. However, for the present, and as a spur to thought (or at least to my own thought), I am posting below a letter about this issue that I wrote to Jared Taylor in 1996. Because I am not good in math, I had to think the issue through on the most basic, beginner’s level, as you will see.

April 5, 1996

Dear Jared:

Here are my notes on the interracial murder rates I discussed with you. The question I was trying to answer for myself was: What is the significance, if any, of your estimate (in “The Color of Crime,” AR June 1994) that the average black is 17 times more likely to kill a white than vice versa? Does this figure really mean what it seems to mean? Does it mean than blacks are 17 times more racist in their choice of victims than whites are?

Let’s start by assuming a population, like our own, with six times as many whites as blacks. For convenience sake let’s say there are six million whites and one million blacks. Further, let’s imagine a situation in which the murder rates by the two groups are equal, say (to pick an arbitrary number) .01 percent. In that case, there would be 600 total murders by whites and 100 total murders by blacks. If the murder victims are chosen randomly by race, then 1/7 of the 600 victims of white killers would be black, or 86 blacks killed by whites, while whites would kill 6/7 of 600, or 514 whites. Meanwhile, blacks would kill 14 blacks and 85 whites. Thus, with equal murder rates by both groups, and with whites six times more numerous than blacks, the average black is six times more likely to kill a white than a white is to kill a black. This differential is purely a function of the smaller number of blacks.

The significance of this is that even though blacks in the hypothetical case are not more violent than whites, the smaller number of blacks means that there are more white victims for the blacks to kill, so it ends up appearing that blacks are more racist than whites, when in fact the differential is only a statistical artifact.

Now, what happens when the murder rates are not equal, but, as in the real world, blacks kill at eight times the rate of whites? We still have 600 murders by whites, but 800 murders by blacks. Whites still kill 86 blacks and 514 whites, while blacks now kill 112 blacks and 681.6 whites. Six million whites killing 86 blacks comes to a white on black murder rate of .0000143. One million blacks killing 681.6 whites comes to a black on white murder rate of .0006816. This works out to the average black being 48 times more likely to kill a white than a white is to kill a black. 48 is also six times eight, i.e., the six-times larger white population times the eight-times greater black murder rate.

Given only the differential in population and in murder rates, individual blacks are 48 times more likely to whites than whites are blacks, which is much greater than the 17 fold difference that is actually the case. Thus blacks are killing whites much less than they would if the murder victims were all chosen at random by race. Therefore, it seems to me that the inference you want your readers to make from the 17 fold difference, i.e., the inference that blacks are far more “racist” than whites, does not hold up.

However, there is a serious problem with all the above calculations—my assumption that in a “normal” population sample, murder victims would be chosen randomly by race. Of course, this is probably far from the case. In a true normal sample, people of all races would probably be much more likely to kill people of their own race. In that case the 17 fold difference in interracial murder would be highly significant. But unless we know what a “normal” interracial murder rate would be, we don’t really know the significance of the 17 fold difference.


Posted by Lawrence Auster at March 26, 2012 03:29 PM | Send

Email entry

Email this entry to:

Your email address:

Message (optional):