The Catalyst

Too many trees have been sacrificed on commentary in the wake of the George Zimmerman-Trayvon Martin verdict. It seems clear that an unconscious racial bias—shared by the perpetrator, the police, the prosecution and the jury—that unfairly associates young black men with violence was partially responsible for this miscarriage of justice. But Florida's Stand Your Ground (SYG) law, which tips the scales of justice in favor of the defendant, especially if there are no credible witnesses, also played a role. If a person is justified in using deadly force “if he or she reasonably believes it is necessary to do so to prevent death or great bodily harm to himself or herself or another,” this law comes perilously close to suggesting that a racial stereotype shared by the culture can be a legitimate defense against a murder charge. Who would have thought that an armed man could attack an unarmed man, and when he starts to lose the fight, shoot him, and successfully claim self-defense? Would a black shooter with a white victim have been acquitted with the same defense?

An interaction (in statistics) occurs when two variables have an effect in combination that is not predictable on the basis of the effects of both of them alone. One type is a catalytic interaction. It occurs when two variables both have the same effect, but their combined influence is much greater than the sum of their individual effects. For example, both alcohol and barbiturates are depressants, but taken together their physiological effect is extremely severe and has resulted in accidental suicides. The one acts as a catalyst for the other.

John Roman of the Urban Institute gathered FBI homicide data from 2005 to 2010 (the last year available), a total of 82,986 cases. The primary variables of interest were the race of the perpetrator and the race of the victim, so unsolved crimes were excluded. The outcome of interest was whether the homicide was ruled justified. Cases involving law enforcement—usually an automatic acquittal—were omitted. Roman also examined whether the case occurred in one of the 23 states having SYG laws. Other control variables available in the data base were the number of perpetrators and victims, whether they were strangers, the weapon used, the year, the region, and the age and gender of the parties. Here are the data:

If we look at all cases, it is clear that both the race of the perpetrator and the race of the victim have significant effects. A homicide is more likely to be declared justified if the shooter is white and if the victim is black. However, the most important effect is a catalytic interaction between these two variables. The shooter is much more likely to be exonerated when a white perpetrator kills a black victim than with the other three combinations, which don't differ very much.

SYG laws also increase the likelihood that homicides will be ruled justified. However, the evidence that they act as a catalyst of racial bias is mixed, since SYG laws increase the number of exonerations in three of the four racial combinations—all but the case when the shooter is black and the victim white.

Would a critic be persuaded by these data? Probably not. It's possible that other variables not recorded in the FBI data base are influencing the outcome, variables such as the location of the incident or the immediately preceding events. A critic might claim, for example, that the white shooter-black victim category includes more home (or business) invasions where standing one's ground is justified. In such cases, we would expect the shooter and the victim to be strangers. As you can see, homicides are more likely to be ruled justified when the perpetrator and the victim are strangers, but lack of acquaintance seems to increase perceived justification in all four racial combinations, not just the white perpetrator-black victim case, as this explanation would suggest. (I drew this conclusion by eyeballing the charts; Roman does not present an analysis of these data.)

Laboratory experiments might help to eliminate some of the ambiguity inherent in the FBI data by creating scenarios which vary the races of the perpetrator and victim and hold other characteristics constant. For example, Birt Duncan showed subjects an ambiguous incident in which one man may or may not have shoved another and asked subjects whether an act of violence occurred. The results were similar to Roman's data; the incident was most likely to be judged violent with a black perpetrator and a white victim. If we are only interested in homicide, we might present participants with written descriptions of killings which vary the races of the shooters and victims and ask them to play the role of jurors. In fact, it wouldn't surprise me if social psychologists around the country are doing that very thing right now.

You may also be interested in reading:

Blaming the Victim

Breathing While Black

Open Carry