Are White Cops Who Shoot Blacks Racist? A Literature Review

With anti-police riots and revenge killings of officers across America, the profession of peacekeeping is under siege. Fueling public outrage is the belief that police shootings of black citizens are driven by racism. Activists see proof of this racism in the fact that while blacks are only 13% of the U.S. population they comprise about 30% of those killed by police. In other words, the percentage of blacks among those killed by police should match the percentage of blacks in the population, unless police are racist. But is that necessarily so? Evidence reviewed in the following video and report suggest not.

Race-of-Officer Research

Instant Access to Current Spot Prices & Interactive Charts

If the higher per-capita killing of blacks by police is caused by racist officers, we would expect white officers to be more likely to shoot blacks compared to black officers. To investigate that possibility I set out to find all research with data on the comparative likelihood of white officers to shoot. After combing databases, references and libraries for months I collected more such studies than any single source cites. This collection, therefore, allows for the most comprehensive literature review to date. And what does it reveal? Among the studies with data for these endpoints:

  • 11 out of 13 studies (85%) find that white officers were less likely to shoot than were black officers.
  • 7 out of 8 studies (88%) find that white officers were less likely to shoot black suspects that were black officers.

Spanning four decades (1981 to 2016), the studies routinely found white police officers were less likely to shoot in general and were less likely to shot blacks in particular compared to black officers. The body race-of-officer research, therefore, fails to support the racism hypothesis for the higher per-capita killing of blacks by police. Moreover, it even suggests that white officers might be holding back from shooting black suspects compared to their non-white colleagues.

For a review of results from each study in the collection, see the supplemental video.

Quality of Studies

As with most studies, the race-of-officer studies typically reply on imperfect data. In general, data on police shootings are far from perfect. However, the results cited above hold across the spectrum of lower- to higher-quality studies. Even when imperfections in some studies are corrected in others, racial bias remains undetected. For example, some studies lacked data on which neighborhoods officers patrolled, yet studies with such data (including one that examined only officers at the same incidents) also found no indication that white officers were more likely to shoot. Such consistency of results across the spectrum of study quality weighs in favor of the results reflecting reality.

Laboratory Research

In addition to that field research, laboratory researchers report similar findings. Using life-like decision-making simulators, Lois James, and colleagues have consistently shown that (1) police officers hesitate fractions of a second longer to shoot armed black suspects than white suspects, and (2) police are more likely to accidentally shoot unarmed white suspects than black suspects. These results also contradict the activist narrative that police shooting of blacks are motived by racism.

Figure 1. Results from James and colleagues showing police officers in laboratory shooting-decision simulators take longer to shoot armed black suspects (left) and are more likely to errantly shoot unarmed whites (right).

In the simulations, police are dressed as if on duty and face life-sized projections of various scenarios involving armed or unarmed individuals. If an officer decides to shoot, she grabs, unholsters, aims and pulls the trigger of her (unloaded) gun just as she would in real life. Yet contrary to activist-directed expectations, when facing lifelike projections of blacks, police do that slower and more carefully.

Implicit-Bias Shooter Tests

The results from those life-like tests differ from implicit bias shooter tests where seated participants facing a computer screen view images of scenes with armed or unarmed people and quickly decide to press a button for a shoot or another button for don’t shoot. A meta-analysis of studies using such tests found that on balance, subjects (1) press the shoot button faster for armed blacks, (2) take longer to press the don’t-shoot button for unarmed blacks, and (3) are equally likely to accidentally shoot unarmed blacks and whites. So if implicit bias studies predict real-life outcomes, we would expect armed blacks to be shot sooner and unarmed blacks to face no greater risk of being shot.

But implicit bias tests may not predict actual outcomes. In the life-like simulations by James and colleagues, even police who demonstrated a pro-black bias in the life-like simulations nevertheless demonstrated anti-black bias on implicit bias tests, subconsciously prejudging blacks as more dangerous. So implicit racial bias may not predict actual explicit responses.

But why then are blacks at greater risk of police shootings?

While police kill numerically more whites than blacks annually, police kill a larger fraction of the smaller black population. But if this higher per-capita killing of blacks is not explained by racist white cops, what is the explanation? To search for an answer we have to broaden our context beyond police and population and consider factors associated with why police shoot, namely crime.

The contribution of a demographic to murder – the maximal violent crime – is a plausible predictor of that demographic’s likelihood of facing lethal police encounters. Indeed, citizens who commit more murders ought to be more likely to face lethal police force. So for example, if men commit more murders, police ought to be more likely to kill men. And unsurprisingly that is the case. So how does the relation of lethal offending to lethal law enforcement hold for the black demographic?

Figure 2. Comparison of four demographics, their share of commissions of homicide on left (source), and their representation among those killed by police on right (source).

On the left of Figure 2, we see the percentages of homicides in the United States by four demographics (the source’s definition of homicide is interchangeable with murder) compared to their percentages among those killed by police on the right. As we can see, for men the anticipated relation holds: the greater contribution of men to homicides corresponds to their greater representation among those killed by police. However, unexpectedly the greater contribution of blacks to homicides corresponds to a lower representation among those killed by police.

So it turns out that while blacks are more likely to be killed by police per capita, they are less likely to be killed by police per homicide. That is also true for women, and we would expect police to be more hesitant to shoot women. So across several lines of evidence we see consistent indications that police might actually be more hesitant to shoot blacks than whites.

In closing

Activists believe that white police officers in the U.S. are engaging in rampant unjustified killings of black men. But how do they know that? Primarily by a handful of cases that go viral on social media and news networks. Yet at the same time, larger numbers of whites shot by police are ignored. Activist knowledge is formed by a relentless selection bias that leads to an unquestioned certitude that had any of the popularized shot black men been white instead, they would not have been shot. However, unless an officer verbalizes racist intentions, we cannot honestly know if racism was a factor even in cases of unjustifiable killings.

Even if no racism existed, some shootings by police will be unjustified, and by chance, some of those will be of black men by white officers. So there is no way to know if a given unjustified shooting of a black person by a white officer was an act of racism. Evidence of racism can only be acquired by confessions of officers or by consistent indications in large pools of aggregated data. Yet after decades of research, there is no consistent body of evidence from large aggregated data sets indicating racist intentions of officers who shoot. What activists believe they know without a shadow of a doubt, in fact, lacks any grounding in the expert literature. Even if at some future date their beliefs acquire sound footing, at this point in time they are in fact but phantoms of the mind.