America is in turmoil because of a video in which a Policeman seems very highly likely to have caused the death of an arrested man. He must face trial. Also, the techniques used in making arrests must be put on trial. Kneeling down on the back of the neck of a hand-cuffed man is not remotely a safe procedure.
A few years ago, Europe was in shock at the picture of a refugee child lying dead on a Greek beach, leading to a chain of events which culminated in the German Chancellor, Angela Merkel, inviting into Germany one million refugees. The picture was more powerful than a thousand words. A thousand times more powerful.
A dreadful image, a horrifying video, can have enormous impact. It brings us face to face with death, from which we are now spared in everyday life, since death has been outsourced to professionals, whereas before it was part of family experience to attend to the dying, and to sit by a visible body in a protracted wake. Equally, the videos of riots, beatings, and shootings cannot help but arouse emotions, of fear, pity, anger and hatred. It is easy to be horrified, repelled by human behaviour, outraged at violence on the part of others, and to wish for prompt and savage retaliation.
Bertrand Russell observed that individual cases had an enormous impact in popular argument. When an individual case is made flesh in the virtual reality of a video, it takes on an incontrovertible status: we saw it happen, as if we were there, and no one can contradict us, other than a liar. We regard ourselves as a witness to a broader reality, and are reluctant to accept that we have witnessed a rare and unrepresentative event. Traumatised victims of crime are apt to over-generalise, both about the perpetrators, the circumstances and place of the event, and about the prevalence of crime. And why not? Their lives were mostly good until that point, and then one or two miscreants put them in fear of their lives, on a street somewhere, in a way they will forever remember. Avoiding those types of people (including their race) and those types of circumstances (talking on a phone in a street) and that particular location seems prudent. The victim has learned from experience, and much therapeutic time may be needed to help them out into public again.
In the face of these emotional realities, from which few are exempt, the search for context may seem a betrayal: the cop-out of the feeble, the excuse of the bystander. Quoting the available statistics may seem perverse, as if it were an attempt to deny that a citizen was killed, as depicted.
Would it be better to show videos of other citizens being killed, in the same or similar manner, to reveal that it happens to men of all races? Bizarrely, this might be a good corrective to over-generalization. Say, the last 20 to 30 cases of death of arrested men, each shown in painful personal detail, might show in virtual reality what comes out of the dry government statistics, that in recent years 3 black men per 10,000 arrests have ended up dead, as compared with 4 white men per 10,000 arrests. @LeonydusJohnson
Will these figures have any impact? I doubt it. We have not seen their faces, heard their pleas, witnessed their agony. It is the difference between a great tragedy onstage and the dry recitation of a telephone directory in a side alley. Statistics are of interest to 0.1% of the population.
Of course, the context should be even broader, which is to look at all racial aspects of crime. These data are not available on an annual basis, and tend to be given as bald numbers, requiring readers to do extra work in order to interpret them. A good starting point is to go through the logic before looking at any numbers.
For each group in society, in this case racial groups, we need to know the crime rate per head of population for all arrestable crimes. This gives us a first approximation, subject to reporting errors, which of course could include differential arrest rates because of racial bias. Where possible, we could compare those arrest rates with victim and witness descriptions of perpetrators. Victims very often see their assailants very close up, and can give racial characterisations, if not descriptions which are detailed enough to secure convictions. Witnesses likewise can generally give racial groupings of perpetrators. Interestingly, self-reports are also available from crime surveys, and these serve as another comparison against which to evaluate arrest rates.
To evaluate the argument that all these figures are shot through with specific racial biases, it would be good to look at speed camera infractions by race.
Crime surveys in which citizens recall crimes (which they may not have taken to the Police) and also give racial descriptions of the perpetrators would be another test case. Interestingly, the argument that the Police and justice systems are biased would have to be extended to all victims, arguing that out of personal racial prejudice the victims mis-identify the actual perpetrator so that the police search for the wrong racial group. This seems unlikely.
Government statistics on race and crime have been published, debated and publicised countless times, yet seem to have no effect at all on the nature of the national discussion. They seem to exist only as a footnote, which should not disturb the main narrative. To mention them is to interrupt an action movie with obscure comments on the possibly misleading effects of selective camera angles.
One reason for confusion is that mainstream media (the BBC is a prime example) give a selective presentation of the findings. This week they were showing graphs (as part of “reality-checking”) of fatal police shootings by ethnicity. What could be wrong with that? What could be wrong with prison population by ethnicity?
The error lies in assuming that racial groups are equally law-abiding. That possibility is not included in the BBC reality checking.
What is to be done?
The statistics don’t matter, and statistics are the distillation of innumerable biographies.