What statistics did they change?
From your cite:
“Control for income” changes the statistics. It’s a measure of the crime rate by income group, not the crime rate.
Just because you can conceive of a situation where the crime **would **be different doesn’t actually make the crime rate different.
Huh? Controlling doesn’t change any statistics. It is an essential part of the hypothesis testing process. If you want to determine the marginal impact of one independent variable on the dependent variable, you control for the rest. [sub]By taking the partial derivative with respect to the variable of interest, for example, in linear regression[/sub]. This isn’t a distortion in any way. No statistics are changed.
The argument is that calculating crime rates by race is not a helpful or interesting way to understand the problem since changing a hypothetical criminal’s race variable does not affect the probability that he will commit a crime.
So abort the poor kids, no matter what color.
They’re saying that when those other factors are the same then the seeming disparity disappears. It doesn’t change the statistics, it just shows that race is not a causal factor in those statistics. They aren’t changing numbers, they’re just controlling samples so that there is equity in the non-racial factors.
I’ve read through this thread and I don’t think that anyone’s pointed out a very important distinction.
I’m pretty sure that babies don’t commit crimes. I’m guessing that you have to be at least, I don’t know, 10 years of age to really be able commit a crime. So you might believe that the current system of arrest and imprisonment is equally fair to all races, and that blacks do make up a disproportionate percent of the prison population because more blacks, on average, have committed a disproportionate amount of crimes relative to other races. I’m not going to argue with you.
What’s racist in Bennett’s remarks is the implicit statement that blacks will ALWAYS commit more crimes. He didn’t say that if you aborted all black babies 20 years ago, our rate of crime would be lower today. That, itself, would be a controversial comment. But you could even give Bennett the benefit of the doubt on a comment like that.
But Bennett said that if you aborted the babies today, it would reduce crime in the future. This is important because it shows that Bennett believes that there there is something inherently criminal within blacks. And that, my friends, is why I think Bennett is a racist.
OK, you don’t like my wording. Change the statistic that we’re talking about. Not change the raw data, but measure a different quantity. “The crime rate” is “the crime rate”. If you want to adjust it for some factor or another, then it isn’t “the crime rate” anymore.
It doesn’t matter that race is not “helpful” in understanding the crime rate. He’s not trying to understand it, per se, he’s just making an observation about what would change it. The fact that you know a BETTER way to change doesn’t mean that Bennett’s observation wouldn’t also change it.
Well, here’s another side to that debate. It’s a long article, and, admittedly is almost ten years old (though only a few years older than the article you linked to), but you would think (or at least hope) that society has if anything become less racist in the intervening years. Some quotes:
It’s not your wording that I object to, it is the validity of Bennett’s observation itself. It is a bad argument.
Your dependent variable is the crime rate. You have a bunch of independent variables to choose from: race, education, income, sex, etc. You want to make the case that the independent variable(s) has marginal impact on the dependent variable.
Bennett observes that lots of crimes are evidently being committed by blacks. He measures only the impact of race on the crime rate. He concludes that being black increases the probability of crime commission and argues to abort black fetuses.
This is wrong. This is a formal proof of why. If you only estimate the impact of race on crime and nothing else, your estimate is biased by an unknown amount, in an unknown direction, and confidence intervals are misreported anticonservatively. For all we know statistically, aborting black fetuses could increase the crime rate.
Omitted variable bias is very serious and is often, as in this case, the product of superficial and highly tendentious thinking.
In fact that is not a proof at all, because when you estimate the marginal effect of a variable you are estimating as if all else was equal. However, the normative statement is still possibly true not because of the causality of black and crime but becuase of the correlation of black and crime. Since race is correlated with presumably the relevant attributes, such as income, by eliminating a certain race you are also eliminating to a large degree a causal factor - income.
I think I hinted at that, but I didn’t go very far into it.
Meanwhile, Bennett is rebuking his rebukers.
I don’t really understand how this relates to what he said. His caller was not proposing a solution where the ends justified the means, so why did Bennett bring up examples of the end justifying the means, including 'abortion of black fetuses would reduce the crime rate?
You make one of the biggest attributional errors that is common in America today. You assume that the end result, incarceration numbers, is indicative of criminality. However, if you would also peruse the other Bureau of Justice Statistics reports, you will find that those who committ crimes, as reported by victims, are much more relative to the racial distribution. Last time my husband had to analyze it was somewhere around 1998 and I think he found that although blacks are incarcerated at a rate of 4x greater than their relative population stipulates, they are actually only responsible for around 26% of crimes reported, which is only 2x. Henceforth, blacks have a 200% increase in their chances to be actually convicted of a crime they commit as compared to whites.
In general deviance reports indicate whites and blacks are very similar, although there are some marked differences in relation to criminal rates relating to drug offenses (especially crack cocaine).
The question our country should be asking is a) why are laws different for offenses predominately committed by lower class inidividuals… i.e. embezzler of 1 million gets 6 months in minimum security, but poor person who grabbed twelve pack of beer and pushed clerk gets 5-15, b) Why are blacks more likely to be prosecuted, found guilty and receive stiffer penalties than whites.
The numbers cannot be denied. The classic stidy of the Saints and the Roughnecks by William Chambliss sheds quite a bit of light on this common assumption.
However, I have little knowledge as compared to my hubby on this matter, since he teaches it in College. I am sure others on this board can back this up more poignantly.
Submitted simply as food for thought…
But all those things being controlled for are not independent of “being Black in America”. Sadly, being Black means you are more likely to be poor. They are only independent in a mathematical sense if you define them to be so.
But we’re not talking about a paper that Bennett is proposing for the American Journal of Crime Statistics, it’s an off-the-cuff remark make in a live talk show. Was he being superfiical? Of course! That’s part of what makes it absurd.
Of course it’s a proof. If you estimate the marginal impact of one variable without other relevant variables, your parameter estimates will be biased in the first place. Controlling for a variable is a fundamentally different thing than leaving it out of the model altogether.
I am not comfortable enough with the data on income distribution to agree with this judgment. Abort all blacks takes out a lot of potential poverty, but it also removes a black working and middle class, which potentially offsets the effects of poverty. Which effect would be larger, I cannot say.
You cannot define variables to be independent if they are not. Even if they do covary, they probably still add new information since they capture different chunks of the variance. If there were a perfect correlation between skin color and poverty (defined binarily), then no new info would be added. This is obviously not the case.
What does “it” mean? 
His remark? Yes, absurd. Indignation? No, perhaps not so absurd.
Bennett says if all black babies were aborted the crime rate would go down. He doesn’t say all blacks are criminals. He doesn’t say there would be no crime. He doesn’t even say the crime rate would be significantly lower. He just said, in effect, 0 Blacks = less crime. Could someone explain to me how this statement is false? I will agree it’s offensive, but it’s not untrue.
Don’t confuse the rate with the absolute level of crime. For the crime rate to go down the removed susbset would have to have a higher crime rate than the average crime rate.
In addition I also like to point out that the incarceration numbers are for those who are imprisoned/convicted for reported, solved crimes.
This is a very small subset of the category “all crime” and a number of factors can significantly skew the results.
Factors include:
-
Person in prison does not equal one crime. So, looking at shear numbers of folks in prison vs. number of other folks in prison does not reflect the number of crimes either are responsible for. In addition, Plea bargains can seriously skew those results as well. So for example, if you had one white person responsible for 1000 crimes, and 4 black people responsible for 100 crimes each, but the incarceration rate would show that the blacks were responsible for “more crime” since there were more of them incarcerated.
-
the numbers incarcerated are incarcerated for solved crime. the “solve rate” is quite abysmal for property offenses as a general category, and not all the comforting for violent offenses. there is not currently a good way to track data wrt racial characteristics of solved vs. unsolved crimes. While it may be comforting to assume that the rates are similar, there’s insufficient data and, more importantly, no control group to assuage the veracity of that extrapolation. There may be other more significant factors in why certain groups get caught.
-
Most importantly, all of this data is for **reported ** crime. This is the easiest way to impact the data in a variety of ways. Consider the crime of shoplifting. It’s only a crime when some one is caught. Otherwise it’s called “shortage”. we simply have no idea how much of that crime exists and the racial characteristics of those committing them.
For a better example, take drug sales and possession. I work in a medium sized city near a Big 10 campus that will remain nameless (but I hear that we’re playing our cross state rivals this Saturday and for the first time in years, we’re ranked and they’re not. But I digress). When Metro Narcotics are on the prowl, they sure aint over at the college campus. They’re in the inner city portions of the city. If they took their dope sniffin’ dog over to the stadium tommorrow, I’m fairly confident they’d be flush with busts. But they aren’t going to be there. And I’m also fairly confident that whole hosts of folks aren’t calling the police to report that they, themselves have been in possession and perhaps sold illegal substances.
In addition, it’s not unheard of for assaults to go unreported, especially domestic styles. BUt in an urban enviornment, it’s easier for neighbors to overhear and call the police, less so, suburban or rural enviornments. What’s the racial characterisitics of those places? see?
So any extrapolation of racial characteristics from the imprisoned population to make the point about racial characteristics of those committing crimes seems to me to be fatally flawed.
I’m not surprised that this thread got posted, as people do like to take shots at stationary targets like Bennett.
I am surprised that it took two hours for someone even to hint at the right answer (cerberus, about context).
I am very surprised that, 13 hours after that, people are still chastising Bennett for anything other than run-of-the-mill insensitivity.
Agreed that conventional crime statistics over-represent crime by blacks, and that Bennett should have referred to sex or poverty instead of race. Nonetheless, the overwhelming popular perception is that the crime stats are accurate, and Bennett was speaking extemperaneously on a talk radio show. He was making a quick-hit point and, as will inevitably happen from time to time, he stuck his foot in his mouth.
So, yes: indignation absurd.
For the record, the book posters are referring to is Freakonomics (now on the best seller lists). One of the chapters in the book examines the dropping crime rates in this country (and others) and comes to the conclusion that legalized abortion and increased numbers of cops on the street helped to make this happen. Many of the other things that have been thrown out as reasons for crime rates to be down don’t hold up when critically examined.
Some points here: There is NO mention of race in the book. The authors state that unwanted children become criminals in a disproportionate amount. This cuts across social class. More abortions = fewer unwanted children = less crime 17-20 years later (when those unwanted kids are most likely to be criminals).
The book has some problems -the self promotion is so thick it practically sticks the pages together, but folks like Bennett no doubt find that a breezy, short, best seller will pretty much do them when it comes to in-depth analysis. It’s important to remember that Freakonomics deals with statistics and epidemiology (which can prove association but no causality).
The addition of the word “black” is all Bennett. And yes, in my mind that makes him a racist. He could have had all the reducto he wanted by sticking to the book and saying unwanted, or any of a host of synonyms.
However, I have to disagree that the Katrina response was racists. It was predominantly classist*. Unfortunately, because so many in the U.S. are completely, totally oblivious to the impact of class in our culture they substitute race, but they aren’t the same. The huge failing of the hurricane evacuation was that (it appears) it never occurred to anyone that there are some people who don’t own their own cars. In reaction to the images coming from NO, the American public was outraged and has responded in unprecedented numbers to provide money, shelter, and housing for the victims. If the response had been primarily racist, the American public would have shrugged and looked away. Rather than being invisible, the NO residents got the lion’s share of the coverage. If this had been a racist response the evacuees would still be sleeping on cots in the Astro Dome while CNN ran non stop coverage of the situation in Biloxi. Every so often they’d do a little feature on New Orleans saying that it was bad there too. Good thing nobody important was affected. While I’ve been embarrassed as an American by all the stupid actions and statements made in the aftermath of the hurricanes, the response of the American public has been extremely heartening.
*There were racist responses and actions, too. I wouldn’t be so naive to think that it was only classist. But the huge systematic failures were ones of failing to recognize and serve the poor and infirm.
Why did he choose black people to make that point?