Are Rookie Cops Less Justified When They Kill Civilians?

Timothy Loehmann was a police officer in his first year of employment with the Cleveland Police Department when he shot and killed 12-year-old Tamir Rice in a park in November of 2014. Peter Liang was a police officer in his first year of employment with the NYPD when he accidentally shot and killed Akai Gurley in a stairwell in November of 2014. These were two of the most well-publicized lethal shooting incidents involving a rookie cop from the last two years, and they might have led one to believe that rookie cops are the problem. Maybe rookie cops have poorer judgment than experienced veterans, and therefore one would expect them to unwisely resort to lethal force more often than experienced veterans.

The Lethal Force Database includes information about the length of time the officer spent on the police force for 768 officers who were involved in shooting incidents from January 2014 through June 2015. Rookie cops were involved in at least 36 lethal shooting incidents. But the highest number of fatal shootings came from officers with six years of experience.

The numbers start to taper off after six years of experience on the police force. This doesn’t necessarily mean that veteran officers are less likely to shoot and kill suspects. It is more likely the case that veteran police officers are less likely to have the sorts of interaction with civilians that would lead to shooting them. Veteran cops become detectives and sergeants and supervisors who take desk jobs.

If rookie cops had poorer judgment than veteran cops, then the average Shooting Justification Rating ought to be lower than the average for more experienced officers. But this isn’t necessarily the case.

It is true that second year cops have a higher average Shooting Justification Rating than rookie cops, and officers with four years of experience have a higher average than either of those two. But after four years on the force, the average Shooting Justification Rating dips lower than even for rookie cops. Beyond about 20 years on the force, the sample size is too low to make meaningful inferences.

This data suggests that the problem with the frequency and quality of justification in officer involved shootings may not be related to experience; that the problem of being too quick on the trigger is not isolated to new officers.  

On Shooting Justification and Race/Ethnicity

According to the Lethaldb.com Shooting Justification Rating, police officers who kill civilians most frequently achieve a rating of 3, which is right in the middle (the scale is 1 to 5). This is the rating for people brandishing handguns or approaching officers with knives.

The average rating is 3.43, on the more justified side of the scale. But it varies by race/ethnicity.

Part of the reason behind developing the Lethaldb.com Shooting Justification Rating was to investigate the differences in the perception of threats presented by white people versus the perception of threats presented by Hispanic and black people. It turns out that Hispanic people, who accounted for only 18% of people killed by a police officer’s use of force between January 2014 and June 2015, accounted for 39% of the killings rated 1 (least justified) in the Shooting Justification Rating. Black people were slightly overrepresented in the two lowest categories of justification as well.

71 of the 1403 fatal shootings of a civilian by a police officer in America between January 2014 and June 2015 categorized by the Shooting Justification Rating were rated 1, 28 of which happened to a Hispanic person and 23 of which happened to a black person. Given the racial composition of the 1403 fatal shootings, a random sample of 71 incidents would include 23 or more black people only 14% of the time, but it would include 28 or more Hispanic people 0.003% of the time. It is possible, therefore, that the overrepresentation of black people in the lowest justification rating is just a random chance. But it is unlikely that the overrepresentation of Hispanic people in the lowest justification rating is random.

What Is This Lethaldb.com Shooting Justification Rating All About?

If you go to my datacard for Michael Brown, you'll see a category called "Lethaldb.com Shooting Justification Rating", with a value in this case set to 2. Every datacard in the Lethal Force Database regarding a person who was shot to death by police has this category. What does this number mean, and how did I come up with it?

That's what this blog post is for.

When I first began investigating police officers’ use of lethal force in 2014, I quickly found that police officers perceived imminent threats that they deemed warranted lethal force from a myriad of suspect behaviors. Often the decedent had fired a gun at or near police, or the decedent had aimed a gun towards police, or at least raised their gun up. Frequently police officers claimed to have felt in imminent danger from a suspect approaching the officer with a knife upraised. Or shootings occurred during physical struggles with officers who felt that they were unable to use less lethal force to stop the suspect from resisting.

But no matter what type of threat was perceived by the officer, they were almost always determined justified by a district attorney or grand jury. Officers who picked fights with suspects or placed themselves in front of suspects’ cars faced the same amount of criminal culpability as officers who bravely faced suspects determined to shoot it out with cops. It seemed to me that criminal culpability wasn’t a strong enough measure to determine which shootings were unjustified in more of a moral sense than a legal sense.

So I came up with a shooting justification rating, and I call it the Lethaldb.com Shooting Justification Rating (clever, I know). The rating is based on a five-point scale, with a rating of 5 indicating the most justified type of fatal shooting and a rating of 1 indicating the least justified type of fatal shooting.  What do I mean by justified? I believe that a police officer who shoots and kills a person with an incredibly lethal weapon (gun) who demonstrates a willingness to use it (is currently firing the gun) has potentially saved lives of both officers and bystanders despite the taking of a life, and the officer’s use of lethal force should be considered highly justified. But an officer who shoots and kills a person without a weapon and who is not confronting officers or bystanders with threatening behavior has only uselessly taken a life for no benefit of the community, and the officer deserves the lowest justification rating possible. But then there are stages in between, like when force is necessary, but gunfire is perhaps excessive, or when a deadly threat is merely potential instead of kinetic.

I’ve done my best to identify the reason that police felt they needed to shoot their guns right at the moment that they did. The Shooting Justification Rating is determined based on the analysis of the incident written by an authority tasked with judicial review of the investigation, like a district attorney. If there is no judicial review or investigation into the shooting, then the Shooting Justification Rating is based on the police department’s narrative of the incident, usually released to the media shortly after the shooting happens. If clear video of the shooting exists and it conflicts with the police narrative, the Shooting Justification Rating is based on the video evidence.

WEAPON: GUN

The Lethaldb.com Shooting Justification Rating ranges from 2 to 5 for decedents carrying real guns, depending on immediacy of the threat level. A point is subtracted in each threat category if the gun turned out to be a replica or non-lethal firearm.

Rating	If the decedent had a gun
5	Decedent was firing gun at officers or others
4	Decedent was aiming a real gun at officers or others; OR decedent was firing a non-lethal firearm like a starter’s pistol, air rifle or BB gun.
3	Decedent was confrontational with officers and holding a real gun; OR decedent was aiming or pointing a non-lethal, replica, or toy firearm
2	Decedent was non-confrontational but holding a real gun; OR decedent was confrontational with officers and holding a non-lethal, replica, or toy firearm; OR decedent had a real gun holstered on his/her person
1	Decedent was non-confrontational but holding a non-lethal, replica, or toy firearm

WEAPON: KNIFE

The Lethaldb.com Shooting Justification Rating ranges from 1 to 4 for decedents carrying knives. This is a step down from the rating for guns because knives are much less lethal than guns.

Rating	If the decedent had a weapon with a blade (knife)
4	Decedent was stabbing an officer or someone else with a knife when shot
3	Decedent was charging towards or lunging at an officer or someone else with a knife when shot
2	Decedent was confrontational with officers and holding a knife
1	Decedent was non-confrontational but holding a knife

WEAPON: VEHICLE

Vehicles are not usually thought of as weapons, but they can be very dangerous to officers in certain circumstances. Officer involved shootings due to suspects in vehicles is highly controversial though in part because bullets can only kill drivers, not stop moving vehicles from becoming dangerous.

Rating	If the decedent was driving a vehicle
4	Decedent was striking an officer (or another) on foot with the vehicle, and the officer had neither jumped in front of the car (thereby knowingly placing himself in jeopardy) nor been able to touch the car when it was stopped (an indication of the speed, and hence lethality, of the vehicle at the moment the officer fired his gun)
3	Decedent was striking an officer on foot, but the officer had either jumped in front of the car or had been able to touch the car when it was stopped (or both); OR decedent was charging toward an officer (or another) on foot with a vehicle, and the officer had neither been able to jump in front of the car nor been able to touch the car when it was stopped
2	Decedent was striking an occupied police vehicle; OR decedent was charging toward an occupied police vehicle; OR decedent was charging toward an officer on foot, but the officer had either jumped in front of the car or had been able to touch the car when it was stopped (or both)
1	Decedent was non-confrontational but driving a vehicle (usually fleeing)

WEAPON: OTHER

Other types of weapons include hammers, clubs, baseball bats, rocks, etc. (usually bludgeoning-type weapons). Because bludgeoning-type weapons are less lethal than knives, the justification for using lethal force against someone striking an officer with, let’s say, a rock is lower than for a knife. However, an officer shooting a suspect charging at or being confrontational while holding an unusual weapon gets the same justification as a suspect holding a knife.

Rating	If the decedent had another type of weapon (usually some kind of bludgeoning weapon)
3	Decedent was striking an officer or someone else with the weapon when shot; OR decedent was charging towards or lunging at an officer or someone else with the weapon when shot
2	Decedent was confrontational with officers and holding the weapon
1	Decedent was non-confrontational but holding a weapon

WEAPON: NONE

Using lethal force against an unarmed person could never be as justified as using lethal force against a person shooting a gun or stabbing with a knife, but an officer facing a suspect using his fists should be as justified in using lethal force as an officer facing a suspect using some other bludgeoning weapon. Besides situations involving actual physical violence, I think there are three other scenarios where shooting an unarmed suspect shouldn’t be completely unjustified.

First, suspects who actually charge at the officer (like Michael Brown), intent on physical harm. The threat is immediate, though the degree is less than lethal. There were 26 of these types of incidents during the 18-month window from January 2014 through June 2015. 

Second, suspects who are reaching out to grab a real firearm. The threat is not yet immediate, but the degree of the threat is quite lethal. There were only eight of these types of incidents during the 18-month window from January 2014 through June 2015, and three of them involved officers believing the suspect was reaching for the officer’s gun. 

Third, suspects who use an object to fool police into thinking they are holding guns. There is no threat to officers, of course, but the shooting was at least not unprompted. There were twenty of these types of incidents from January 2014 through June 2015.

Rating	If the decedent was unarmed
3	Decedent was in a physical struggle with a police officer or someone else
2	Decedent was charging at a police officer (or someone else); OR decedent was in close proximity to a real firearm and was reaching for it; OR decedent was holding a non-weapon in a threatening way to try to prompt a police response
1	Decedent was in close proximity to a replica, toy or BB gun and was reaching for it; OR decedent was non-compliant with orders but not physically threatening to officers or others; OR decedent was compliant with orders and also not physically threatening to officers or others

Every other situation gets a justification rating of 1. Officers often shoot unarmed people because they are unable to see the suspect’s hands and believe the hands may be concealing a gun. But in my opinion, the uncertainty of a threat should not justify the use of force intent on killing a person, even with 20/20 hindsight. 

The Lethaldb.com Shooting Justification Rating is merely a subjective rating, but I was able to apply it to the vast majority of lethal shooting incidents (1403 of 1406 incidents from January 2014 through June 2015). Here are two examples of the kinds of analysis that can be performed using this data.

Lethal Force in the Rural West

The rate of fatal incidents involving law enforcement officers is higher in Western states than in the rest of America, and policing in rural areas has a lot to do with it. This is the argument made by Kate Schimel in an article written for the High Country News and republished in the Denver Post on Saturday. And it’s true that Western states see more killings by law enforcement per capita per year than most of the rest of the country (Oklahoma excepted).  Of the five states with the highest numbers of lethal force incidents between January 1, 2014 and June 30, 2015, four of them are Western states (New Mexico, Arizona, Nevada, Montana, ranked 2, 3, 4, and 5, respectively. High Country News names these four, plus California, Oregon, Washington, Idaho, Utah, Wyoming and Colorado, as Western states).

Rate of lethal force incidents per million per year by state, January 2014 through June 2015

But this doesn’t say much about the how often people in rural areas get killed by a police officer’s use of force. Schimel used the story of Idaho rancher Jack Yantis in her article as an anecdote of a rural officer-involved shooting. Yantis was shot and killed by Adams County sheriff’s deputies after they alerted Yantis to a situation with a bull of his that had been injured after it wandered onto the highway. Yantis showed up with a shotgun in order to shoot the injured bull, but according to bystanders, words (and possibly bullets) were exchanged between the deputies and Yantis.

But cattle-based officer-involved shootings are very rare, and they are not the reason that officer-involved shootings happen at a higher per capita rate in the west than elsewhere in America.  While stating that “rural areas have been equally bloody” as cities in the West, Schimel acknowledges that “the lack of comprehensive long-term data makes it difficult to know how the rate of rural fatalities compares to urban zones, but their frequency is notable.” We may not have long term data, but as part of the Lethal Force Database, we do know the rates of lethal force incidents (LFI) in each metropolitan statistical area (MSA) in the time frame examined, currently January 2014 through June 2015.

A few notes about this table: first, the metro population listed for each state is the sum of the population within all metropolitan statistical areas with a central city based within that state, not for all parts of the state within a metropolitan statistical area. This makes a difference in the east, where metro areas frequently cross state lines, but there are only a handful of metro areas that cross state lines in the west (Portland, Oregon; Logan, Utah; Lewiston, Idaho). Second, metropolitan statistical areas coincide with county boundaries, and often include land that most people would consider rural. Similarly, areas outside of metropolitan areas contain many small cities (the threshold for a city to be considered a metropolitan area is about 100,000 people) whose police forces interact with the public in the same ways as mid-sized cities.  This is all a way of saying that this table may not reveal answers about the rural American police force in quite as clear cut of a way as it is presented due to certain issues central to the structure of census-defined metropolitan statistical areas.

The idea presented in Schimel’s article is that the rural areas of eleven Western states may be plagued by as many (and perhaps more) fatal acts of force by police as the attention-grabbing incidents in Ferguson, Chicago, North Charleston and elsewhere. This idea is partially borne out by the data. The difference between the non-metropolitan rate of lethal force per million per year and the metropolitan rate of lethal force was largest in Arizona, where nine people have been killed by police in Arizona’s non-metropolitan counties, for an annual rate of 12.5 people killed by police per million population. Three out of the top four states ranked by this difference are Western states (Arizona, Colorado and Idaho). Only eighteen states had rural rates of lethal force higher than urban rates (defined by metropolitan statistical areas), and six of them were Western states. But five Western states saw urban rates of lethal force greater than the rate in rural areas (Oregon, Nevada, Utah, New Mexico and Montana).

According to Schimel, only two out of the 21 lethal force incidents in the state of Washington in 2015 were committed by Seattle police officers. This was seen as evidence that rural areas must be responsible for the rest. But this isn’t actually the case. From January 2014 through June 2015, the annual rate was about 3 people killed per million residents of the Seattle metro area. (For reference, the national average during this time was 3.17 deaths per million population per year.) Non-metropolitan Washington experienced a rate of 4.3 deaths per million, a rate modestly higher than Seattle’s. But the small metropolitan areas of Yakima and the tri-cities of Richland, Kennewick and Pasco experienced a rate of 13.6 and 14.8 deaths per million, more than four times higher than Seattle’s rate and more than three times higher than the state’s non-metropolitan rate.

Indeed, the smallest metropolitan statistical areas around the nation account for the highest rates of lethal force incidents. While very large metropolitan statistical areas (New York, Los Angeles, Chicago, Houston, etc.) account for substantially fewer officer-involved deaths per capita than the rest of the country, the rate of lethal force per capita in rural areas is also less than the national average.

There are a few different explanations for this. One is simply that non-metropolitan areas have less crime per capita than metropolitan areas, and fewer criminals means fewer chances that a person with a weapon gets killed by police. Another theory is that the database is simply missing people killed by police in rural areas.  The incidents in the Lethal Force Database were gathered from sources that find media reports about officer involved shootings, tasings, and physical force that result in fatalities. There is no guarantee that this is a complete list, and it is possible that small town media slips through the algorithms (small newspapers are more likely to have paywalls and non-searchable archives, I’ve found).

What is definitely true is that urban areas of Western states lead the nation in lethal force incidents. But crime in rural areas seems to be on the rise, and drug overdose deaths in rural areas are becoming more plentiful. While there is conflicting evidence of higher rates of lethal force incidents in rural areas, long term studies may show us that more crime and drugs will bring more officer involved shootings to America’s rural communities.

Unarmed People Killed by Police

Prior to 2014, the official word on the number of homicides committed by police officers in the line of duty came from the FBI and their annual reports on crime statistics. But thanks to the work of the people who run databases like killedbypolice.net and Fatal Encounters, it was shown that the FBI was vastly undercounting the number of people killed by police officers in the line of duty. This was once again the case in 2014 when the official count from the FBI showed 444 people dead from officers in the line of duty[1].

But the true number, according to my research, was 1,004, including 939 deaths from officer involved shootings, 39 taser-related deaths, and 26 deaths from other means.

Only 55% of police killings from January 2014 through June 2015 involved people armed with a gun. 18% of people killed by police were unarmed.

Even when isolating the killings that occurred when police officers used lethal force (guns), one out of every eight people (12%) were unarmed at the time they were shot.

Black Lives Matter activists have done a good job of bringing attention to police shootings of unarmed black men, and this has reinforced the belief that police officers disproportionately target unarmed black men. There is some truth to this. Even though 26% of people who died from police gunfire were black, 35% of unarmed people who died from police gunfire were black. Including deaths from tasers and physical struggles, black people made up 39% of all civilian deaths from police officers from January 2014 to June 2015.

But the circumstances by which police end up shooting unarmed people are not similar. Michael Brown was unarmed when he was killed by Ferguson police officer Darren Wilson, but he was charging threateningly at Wilson when Wilson pulled the trigger. Ezell Ford was unarmed when he was killed by Los Angeles police officers Sharlton Wampler and Antonio Villegas, but he was also allegedly in a physical struggle with the officers when they shot him. These officers responded to a threat of physical harm by using lethal force. Shooting an unarmed person in a fistfight is a disproportionate response, of course, but at least the threat in these instances was real and not perceived.  Other unarmed people held objects (cell phones, remote controls, power drills, etc.) in order to fool police into shooting them. In these instances, the threat to officers was merely perceived, but good arguments have been made saying that a reasonable officer should not be found criminally culpable for believing the decedent in these types of situations had a weapon.

However, 55 people from January 2014 through June 2015 were killed by police even though they had no weapon, real or perceived, and posed no physical threat to officers or bystanders. Most of these types of shootings happen because police officers are unable to see the hands of the suspect they are pursuing. But in these shootings, police officers decided to shoot the suspect first rather than wait for a suspect to present a threat either to officers or bystanders.

This sort of quick response is what officers train for, and they are often commended for quick action by their superiors. But by removing the idea that a threat must be imminent before lethal force is used, it allows police officers the right to deprive a person of life without due process of law.

While it is true that these 55 police shootings made up only a tiny percentage of the total number of shootings in this time frame (specifically 4%), the shootings happened often enough to be able to make statistical inferences from them. And what I found was intriguing.

While black people made up 26% of the total police shooting deaths from January 2014 through June 2015, black people who were unarmed and not a threat were 31% of the total number of people who were both unarmed and not a threat (17 of 55). This is only a modest increase, however, and it is possible that this could just be due to random chance. If a random group of 55 people were selected from the 1,403 people who lost their lives due to a police shooting from January 2014 through June 2015, the probability that 31% or more would be black is about 1 in 4. This would be an unlikely result, but not an uncommon one.

But, despite making up only 18% of total police shooting deaths from January 2014 to June 2015, Hispanic people made up 40% of the police shooting deaths (22 out of 55) of people who were unarmed and posed no threat.  If a random group of 55 people were selected from the 1,403 people who were shot and killed by police from January 2014 through June 2015, the probability that 40% or more would be Hispanic is 0.02% or about 2 in 10,000.

It seems highly unlikely therefore that this result is due to random chance. But then that gets to the question of why. Why is it that Hispanic people are more often the victims of fatal police shootings of unarmed non-threatening people than any other race or ethnicity? It could be that police departments with loose rules governing their use of force policies also happen to be located in areas with high populations of Hispanic people. It could still be a crazy statistical anomaly, and that data from the next 18 months will show that it was just a fluke.

Or it could be that officers are more likely to believe that Hispanic people have weapons when they really don’t. This would be highly problematic if true.

---

https://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2014/crime-in-the-u.s.-2014/tables/expanded-homicide-data/expanded_homicide_data_table_14_justifiable_homicide_by_weapon_law_enforcement_2010-2014.xls

Phoenix Police Spokesman Misleads the Public About Phoenix Officer Involved Shootings

When a police officer uses lethal force in Phoenix, the Phoenix Police Department's spokesman Sergeant Trent Crump is the person in charge of giving information to the media. Sergeant Crump was asked about a study commissioned by Phoenix PD, performed by Arizona State University researchers and released in October 2015.  In the statement he gave to the media, Sergeant Crump claimed that Phoenix was "one of the safest major metropolitan cities in the United States when it comes to officer-involved shootings."  Crump cited the number of police shootings in 2014 and compared it with numbers of other (larger) cities: "Los Angeles had 30, ... Chicago had 45, Houston had 35, Philly had 29, Phoenix Police had 21."


But it's not that clear that Phoenix is one of the safest cities for avoiding officer-involved shootings.   Though Phoenix had fewer officer involved shootings than every other city on this list, their shootings were deadlier than other cities.  Los Angeles and Phoenix were the only two cities where greater than 50% of OIS's ended up in the death of the target. Though Phoenix police officers shot at only 21 people in 2014, 16 people ended up dead due to a Phoenix police officer's use of force, including 13 people who were shot and killed.



It gets even worse for Phoenix when population is factored in. Phoenix goes from first to a distant second in terms of rate of OIS's per million residents. Phoenix police officers killed people at a rate much higher than officers in Los Angeles, Chicago, Houston or Philadelphia.



Among the 55 largest cities in the United States, Phoenix PD was seventh in terms of the rate of lethal force events per million residents. (The chart below uses 2013 population estimates.)



It could just be a fluke that Phoenix police were so deadly in 2014. After all, only four officer involved shootings resulted in fatalities in the first half of 2015.

Data from lethaldb.silk.co

But there's reason to believe that the Phoenix police is actually one of the deadliest departments in the United States.  The Arizona State University found that 64 people had died in the six-year period that it covered, 2009-2014. This would be a rate of 6.9 people killed per million per year. In 2014, this would have placed Phoenix 15th on the list of 55 largest cities by population. This would still rank above Chicago (19th), Houston (25th), Los Angeles (27th) and Philadelphia (41st).