Hi I’m Wendy Zukerman and you’re listening to Science Vs from Gimlet. This is the show that pits facts against force. On today’s show: when police kill people. Heads up this episode has some violent descriptions and swear words. So take care when you’re listening.

It’s been five years since the high profile shootings of several unarmed black teenagers and men[1]...which launched the Black Lives Matter movement[2]...  

Black Lives Matter! Black Lives Matter!

Thousands of people protested saying... cops are using unnecessary force --

<<I can’t breathe! I can’t breathe!>>

The deaths sparked national debate about what exactly was going on in Police Departments…with some saying this is a systemic problem … and other saying no, that’s not right. A recent Pew Survey found that around 40% of the public and about two thirds of officers interviewed that thought these deaths were just quote “isolated incidents”[3]

<<This whole anti-police rhetoric is based on a lie. there is no data, there is no research that proves any of that nonsense. None. >>

But despite this, police departments have been doing all kinds of things to respond to the deaths and protests.

Police departments are employing a new tactic: Body mounted video cameras

The Sacramento pd has been doing implicit bias training

Officers and recruits are required to attend an 8 hr implicit bias course-04

Meanwhile,  it feels like we’re still seeing a lot of black people being killed at the hands of police just a few weeks ago…

Put your hands up! Show me your hands!

Atatiana Jefferson was shot in her home in Texas..

Jefferson had been playing video with her 8 year old nephew … When she was gunned down in front of him

So,  today on the show: What can new research tell us about police shootings in the US? And what can police departments be doing to save lives? Do implicit bias training - or body cams actually help? 

When it comes to police shootings there are a lot of opinions… But then there’s science…


Ok so when you want to know the science of what is happening with all these police shootings… you have to know. We don’t have good data here. There is no official federal database that tallies all the people in the US who were shot and killed by police[4][5].  

So without official stats[6] , media organisations[7]  like the Washington Post[8], have started counting. Here’s what they’ve found. About 1000 people in the US are killed by police each year. Roughly half of those don’t have a gun[9].  And while white[10] people do get killed by cops - you’re two-three times more likely to be killed by police in the US if you’re black.[11][12]

And every researcher that we spoke to told us this wasn’t just on the cops. So for example, this difference is partly explained by the fact that - because of the history of this country - black people are more likely to live in places with gun violence.[13][14][15]  And cops are more likely to be assigned to these neighbourhoods[16][17]. So one idea here is that if a lot of cops are interacting with more black people than white people - in the rare times cops do draw their weapon - there’s a greater chance the person on the other end of the barrell is black.[18] [19] But lots of people in the US think someone else is going on too… racial bias - or racism - among cops. So how widespread is that? Well..  what feels like a messy and politically charged question… can actually be answered with science.

JE Hi, My name is Jennifer Eberhardt and I work at stanford University.[20].  

Jennifer’s a professor of social psychology And in 2014 she started working with the Oakland Police Department[21][22] to analyze their data on police stops and searches. And she was basically asking the question: that when you look at the data: are cops stopping and searching more black people? She looked at more than 28,000 stops from 500 police officers[23]. And  she said there was a clear pattern:

JE: They stopped way more African Americans and the same was true for the searches.

WZ: And when they did the searches, what did they find, were black people more likely to have weapons or drugs or something like that?

JE: No no they weren’t.

In Oakland, black men were searched four times more often than white men… despite the fact they were no more likely to have weapons or drugs than white guys.[24] A similar trend was found for latinos.[25][26][27]  


And this isn’t just Oakland… One big study[28] published just this year analyzed almost a 100 million police stops from 28 different states[29] all across the US and they  found similar patterns.[30] [31] and that these differences often stick around even after you adjust for crime rates. [32] [33] [34]

And so Jennifer wanted to know what exactly was happening in the interactions between police and the people they stopped. Like, besides being stopped, how were black people being treated?  And she found a cool way to study it.[35] Body cams.

JE For the first time we have these body worn cameras and we have footage were ce can actually see - We can be actually at the scene right when this is happening. And as a social psychologist. This was just like gold right.  

Oakland PD gave Jennifer their body cam footage from traffic stops - y’know when a cop stops someone for speeding or having a broken tail light.[36]  And Jennifer’s team transcribed the conversations between cops and drivers.. From nearly 1000 traffic stops. They took a sample, and showed the transcripts to people waiting in line at the DMV - easily the most exciting thing to happen in a DMV in years[37].

So they’re Oakland drivers, for the most part, and we had them rate the officers language. Yknow, they didn’t know the race of the driver at all the, they would just look at a transcript.

And she got them to rate the conversation - to see how respectful and fair the officers was being. Now, overall, people thought the officers were generally professional. But

JE: There was a difference in how they spoke to white and black drivers.So they spoke to black drivers with less respect than white drivers and you could see this at the very beginning,

It started in the first 5 seconds.[38]

JE So this was before the driver could say a word So with white drivers they were more likely to greet them with sir or ma'am or mister.

But with black drivers?

JE They use words like Bro... dude.  

WZ Bro? Oh my gosh

JE Yeah. Yeah.

The cops were also more likely to tell black drivers to “keep your hands on the wheel,” and were less likely to reassure the black drivers... [39].

JE so they would say to white drivers things like you know it'll be OK. You know don't worry you know that kind of thing. Which they were less inclined to do with black drivers.

WZ I'm literally just thinking about all the times that I've been stopped by a cop and I think they have always said nothing to worry about here. I just got to ask you some questions. And then like Be safe out there or something you know you always get that Oh I always get that. Be safe out there.

Jennifer is black - by the way.

JE I also I have never been told by an officer that I was stopped by nothing to worry about here.

Hahha… I have to say.

She thought maybe -- it was just a few cops behaving really badly -- but another look showed no. It was happening across the board.[40] Even if the cop was black or white --on average they all treated black drivers worse  the differences were stark.  Jennifer actually fed all her data into a machine learning algorithm. And just analyzing what the cop said…  the algorithm could figure out the race of the driver…[41] 

JE based on the words the officers used alone, we could predict whether that officer was talking to a black person or a white person.

WZ whoa

JE So it's a real real effects here.  

WZ: When the computer program spat out the answer and got it right a lot of the time, were you surprised?

Were you surprised not- mildly, yeah i don't know, crystalized everything, put it that way..

Jennifer looked at a lot of different things to try to explain the data… like, maybe black drivers were getting stopped for driving more dangerously. So then cops came on strong? No no…that wasn’t it. She looked at gender, criminal history, location of the stop…

That didn’t explain it. Everything we considered it didn't make much of a difference-01

But of course there is one obvious explanation here. Cops are intentionally treating black people with less respect or - maybe they’re behaving differently due to some kind of unconscious or implicit bias that they have.[42][43] Jennifer thinks could be part of what’s going on.

JE: So I think when people think about racism they’re thinking about people filled with hate, and they’re thinking about people who burn crosses… for implicit or unconscious bias it can be a bias even when you don’t have it and…..  even when you don’t have a bad heart.

Scientists study this kind of bias using a test called the Implicit Association test. In one version, you see pictures of black and white faces alongside weapons or harmless objects. And the test basically records how you react. It’s not a perfect test - but it finds that more than 80% of police officers tested link black faces to dangerous weapons [44].. It’s not just cops who have implicit bias - a lot of us do.[45][46][47] 

It’s thought we get these associations from things around us -  movies and TV.  And these associations can happen pretty quickly… Jennifer told us a story about one cop she spoke to who was new to the US and he told Jennifer could feel these biases kicking in.

JE He wasn’t thinking about a black person in terms of fear and threat and aggression. But when he came to the US he was, and a lot of that he felt like came from the police work he did he would hear black man, black man blaring on the radio. And so he felt like, after all while he could feel it affecting him when he wasn’t working, and when he saw a black man he’d look for his hands and be on high alert. And he said even his friends noticed that he was changing- and that he was doing this And his friends would call him out on it - and he could tell he had changed!  And he was worried about it. He could see- feel this bias creeping in to how he thought and how he looked at black people.

So there’s evidence cops are treating black people in America different to white people. While we don't have great data on police shootings, it’s clear in the data on stops, searches... And in Jennifer’s study it was also clear when you looked at how police just said hello. So the big question right now. Is how do we fix this? Are there any scientifically backed solutions out there?

Let’s first take a look at Implicit Bias Training. If we know cops have implicit bias - the maybe the answer is giving them training to fix it?



In the past decade, implicit bias training has become all the rage among Police Departments[48][49][50]. Programs have popped up all around the country from New York[51] to  Kentucky[52], to Oregon[53][54]  The Department of Justice has gone mad for them too. A couple years ago[55][56] they announced they’re officially getting all federal law enforcement officers and prosecutors to do this kind of training. And it’s a big deal. In some cases these workshops can go for several days… and can cost hundreds of thousands of dollars for a police department[57][58][59],  so we wanted to know, do they work?

For this we talked to Professor Phillip Atiba Goff who runs the Center for Policing Equity which based at the John Jay College of Criminal Justice in New York[60]  … He works with police departments to prevent unnecessary force, he says over the last few years, police chiefs are calling him up left right and centre.

PG they came out of the woodwork.

WZ police chiefs?

PG police chiefs, yes.

WZ they were calling you to say help?

PG Yes “please, help us make this better.” This being: too much force, racially disparate force,, help out make our department less deadly, more fair. Came out of the woodwork asking for it. Black my whole life, did not see this coming.

And one thing that lots of police chiefs ask Phil is implicit bias training. So Phil told us that in these implicit bias workshops a big thing is just to teach cops: what is implicit bias. And getting them acknowledge - that like a lot of us - they have them too. And Phil says that isn't always easy.  sometimes they won’t cop to it.

PG: Some [police] folks think that this is BS and it's a way to call them bigots,  people don't understand policing

But Phil says there are ways to talk to cops to help them see implicit bias differently… likes he told us this story of a cop who was called to a shooting. The first thing the officer thought? Oh, this guy’ll be black, maybe latino[61]

PG So he walks in, sees a woman and he’s like I have to get this woman behind him and out. But as he tries to make that move to get her behind him And then he realises she’s not panicked she’s not scared, she’s exactly where she wants to be. This is going to be one of the shooters. So he draws on her at the exact same time that she shoots on him. 

The cop survived, the shooter was taken to hospital. And for Phil, this shows the dangers of falling prey to your implicit bias.

The implicit bias there was that shooter means male. And in this case there is a male and a female, man and woman and if you rely on your implicit biases, that officer - who lived to tell the tale, would have been dead. They understand it. And as soon as it clicks for them that it’s not just gender, it’s race, and class, neighborhood. They understand that failing to recognize it is just a survival thing.

So getting cops to understand implicit bias. That’s one thing. But it’s hoped that these trainings will also reduce that bias. And there’s some evidence that they can help.[62][63][64][65] One big meta-analysis[66] looked at implicit bias trainings for all sorts of people- not just cops. It showed trainings can reduce people’s bias. But it’s not like they magically erase it. Like often the effect doesn’t last long. After a few days  - people go back to their old ways.[67]  And when it comes to cops specifically - there’s hardly any research on whether these trainings change police behaviour, like reduce use of force. Which is why Phil says…

I am not yet convinced that the marketplace of people who charge for implicit bias training that that’s a valuable add to law enforcement… >>  

On top of this - Philip told us that because of the demand for these training sessions, all sorts of people end up running them. No one is policing this.


<<PG: There are folks who are not scientists[68] That doesn’t feel comfortable to me and the rush to do something on implicit because we know it’s a real thing in the world, doesn’t mean the trainings are doing what we want them to be doing.

And the stakes are high - because it’s possible that crappy implicit bias training can back-fire.[69][70] Studies have found that showing people these biases - can actually validate them and people end up stereotyping more[71][72]. 

Ok. So Implicit Bias Training. Not a slam dunk. The next idea we want to interrogate are body cams[73][74]. Around 60% of police departments in the US use body-cameras[75]. 60% And some are hopeful about this - because they think that if cops are recorded all the time. They won’t behave badly. So do body cams prevent police shootings, or other unnecessary force?

One of the best studies we found on this[76]  was was run by David Yokum from Brown[77]. He’d heard a lot of hope about these cameras.

It wasn’t this is one or two instances that this is going to make a difference, it was this is going to dramatically curb uses of force and complaints across the board.

To see if that was true… David zoomed on Washington DC. That police department was about to roll out it’s Body Cams program. And so David thought. Hold up! This could be a perfect opportunity to study this.. And the police department agreed to play ball. So, instead of giving all the cops the cameras at once - David thought let’s randomly give some officers body cameras, and leave others without them

we'd kind of flip a coin, heads they'd get a camera, tails not a camera

And they were flipping a lot of coins…this was a big study

Just a little bit over 2200 officers,

Oh wow.

So it's a very large study the largest in the world

David’s followed these officers for about 7 months[78]. Tracking citizen complaints against officers. And uses of force -- anytime a cop tackles someone to the ground or draws a weapon,[79]. 

DY Did use of force go up or down, did they stay the same?

WZ did you have a hunch on what was going to happen?

DY I really didn’t. I didn’t.

So what did he find?

DY The groups of officer with cameras and without cameras look to have the same uses of force.  We didn’t actually detect any meaningful differences on anything

WZ: On anything?

On anything? 

A review paper - looking at 10 other studies on body cams- found the same thing[80]. They had no effect on police using force. Now, studies like David’s can’t tell if body cams will affect shootings specifically because they are so rare. You would need to track things for more than 7 months to get proper data on this. But still, David told us that he doesn’t think body cams would have a big effect on shootings either. After all - he says - there are quite a few cases where police were being filmed shooting someone unnecessarily[81][82].

When you give examples like those Youtube clips, where everyone has a camera and you think - why didn't that change behaviour in those movements? Like if you slow down and think about it’s maybe not obvious as it felt years ago.

But David told us that even though these cameras might not reduce force overall … they may be useful in very specific circumstances. Like for people to know exactly what happens when something does go wrong.

if I was the emperor of the day, I would not immediately deactivate the program, but a different road If I was in a jurisdiction that didn’t have a camera program, I would not buy them right away

 So we’ve got one big problem. And two strikes out. After the break… things get more complicated… but then a breakthrough..

PG I’m credibly optimistic about where we are in policing I think it’s the best news race story in the country…  


Welcome back. We're looking for a way to cut down on unnecessary police violence. It feels like an intractable problem… But after calling scores of academics and police chiefs… we found something that we think can help. But to really understand why it might work. We need to take a closer look at what’s happening in the intense situations that can lead to someone getting shot by police.

So I went to a shooting simulator in Spo -can, Washington.  With my producer Meryl Horn. Cops often get trained here...

LJ I need to ask everyone: any live weapons?

MH No WZ? No.

LJ Go on through

That’s Lois James[83], a researcher at Washington State University. We walked into a small room, about the size of a squash court…

MH So we’re in the simulator…

On one end of this room, is was a giant video projection screen… that’s where Lois plays videos  she’s created. They’re based on real situations that police have been in, which have lead to shootings.

LJ  I looked at 30 years of data collected by the FBI on a typical officer involved shootings.

A lot of police shootings started with a call for a “domestic disturbances” - arguments between family members or roommates.[84] So to see what that’s like… I strapped on a holster and was handed a gun. Lois’s husband Steve James[85] is a researcher too. He runs the lab.

Steve: so there’s nothing in the magazine, and we’ve screwed a laser into the glock. So it fires that when you squeeze the trigger.

WZ So I just need to take it out of the holster, and pull the trigger.

And heads up there’s some violence in this next 40 seconds. If you’d rather skip ahead, do it now.

The simulation started, I walked into a house… and a couple was fighting… I walked down a hallway… and was looking down a staircase.

Guy: you are not leaving this house. NO! You are not leaving this fucking house this is my baby- did you fucking call the cops?

WZ: police here. Police here Quiet down please.

It was kind of hard to see what was going on… I wasn’t sure if the man had a gun. Lois told me it’s important to look at a suspect's hands in case they have a weapon

WZ Can you please step aside can you show me your hands, show me your hands. Show me!

(Audio running under) Then, he pulled out a gun and started shooting at me. And I shot back.


Bang bang. Baby crying.

After it was over, Steve showed me what happened…

SJ Look how quickly that was, you fired 6 rounds in 2.5 seconds.

WZ Wait so after I shot him once, I just kept shooting?

SJ Yeah- you hit him 3 times in the chest.

I didn't realise how quickly things can escalate in these situations…. Everything happened so fast. And Steve said it’s not just podcast hosts who have these reactions.

We've seen officers triple their heart rate even though the threat isn't real.

Studies of cops, trainees and soldiers[86][87][88] [89] have found that in the heat of the moment, the body can be flooded with stress hormones, and this is when it’s more likely that they’ll make mistakes.[90][91][92] Like, in other scenarios that Lois runs…  instead of a gun, someone whips out a black wallet - and a few cops got confused and still shot. And Lois thinks this might be happening in some of the situations we see in videos where police kill unarmed people for seemingly no reason.

LJ In these situations where it’s like how in the world could the officer have shot in that situation and there’s footage of it, where it goes horribly wrong and it’s clearly the fault of the office, I would speculate or certainly argue it’s fear based

WZ: What do you mean by that?

LJ: The officer is not thinking, they’ve lost the ability, they’re reacting and not thinking.

What Lois and Steve’s simulator tells us. Is that if it can be very difficult to stop what’s going on once you’re in the heat of the moment… So we need a solution - that kicks in sooner - before things ever getting this bad.

To find out how this might be possible, we went back to Professor Phillip Goff who you heard from before talking about implicit bias training. He’s got an idea on that seems to be working.

PG: So

WZ hold on- siren

PG that’s not just tone painting for what we’re doing?

WZ we’ll obviously have it running through

PG I would assume, yes

Ok, let’s get our own siren goingSo Phil told us what he thinks gives us the best chance to stop unnecessary force in policing…Ok ok.. Cut the siren. Phil says we shouldn’t focus on fixing what’s in the minds of police officers… Instead, he says. Make rules.[93]. and policies that stop police officers getting into unnecessary situations where someone might get killed.

PG I don't care that much about saving the souls of people who may or may not be engaged in racial discriminatory behaviours, I care about your behaviors. Solving the hearts and minds thing is a distraction from fixing the behaviours… So I’d much rather change policy than change your mind

And he says that finding out what exactly what policies need to be changed… might be different

for each city. As an example, Phil told us about happened when he got a call from the Las Vegas Police Department. There had been some high profile shootings of unarmed men, and calls for reform were growing louder.[94][95][96] So Phil wanted to dive into their data to find out what needed fixing. The department was like - great. Let’s do this. Only problem?

PG There were rules in Nevada about sending data from a police department.

So, Phil drives out to the desert to collect the files. And he figured heck let’s have some fun with this?

PG I was wearing a black double breasted thick white pin striped suit with black hat, and red carnation. It was very hot, since it was the most cloak and dagger thing I could imagine. We rented a car, they rented a car, 2 black cars, we met in the desert, they opened their trunk, took out a box of papers, they put it in our truck.

[TRACK NOIR SOUNDING] Phil went back to the office - presumably still wearing his pin-striped suit. And the started sifting through the data. And he was looking for patterns when cops tended to use a lot of force.

PG Was going through the files, saw foot pursuit foot pursuit foot pursuit. That’s a lot of foot pursuits.

Foot pursuits are basically when a cop has to run after someone running away from them. And Phil was finding that in Las Vegas - there were escalating really quickly in these pursuits. That’s when police tended to use a lot of force. And Phil totally gets why. He’s been on a few ride-alongs with cops where he was in a foot pursuit.  And he said that by the time they caught up to the suspect, even nerdy pinstripe wearing Phil was pissed off! He was ready to beat up someone.


PG Like, my politics are to the left of Ghandi, and I wanted to give him a shot to the kidneys. Cause you’re hopped up on adrenalin, you’re like this bad guy is putting me in danger! Come here!

So Phil told the chief. Here’s the rule. To help stop police force. Tell your officers: when you’re in a foot pursuit and you catch up to the suspect, say they’re surrounded by cops[97] and has stopped running. Slow down. The new training suggested they count to 10 before doing anything.

PG So before - you’re on a foot pursuit guy is like please don’t hurt me, tackle, now- I'm going to hurt you- 1 2 3 4 5. Oh i really want to beat you up. 7, 8, 9, 10. Hands on your head. But in general it was slow down- make sure you’re responding to the current situation and not just to your adrenaline.

After a couple years of these new kinds of policies, the Las Vegas PD checked to see if this worked.

PG when they started training their officers to do that, the following year they saw a 23% reduction in police use of force across the board.[98]

23%- so almost a quarter less use of force incidents - including large drops in the use of pepper spray, tasers and handcuffs.[99] A lieutenant from Las Vegas PD came to work for Phil… and he told them…

Hey you know that stuff you guys did? It really worked! And that was exciting to us. And we thought! Hey! That stuff we thought worked? It worked! 

Phil’s ideas are part of a larger strategy that’s sometimes called “De-escalation”.[100] And it’s basically bringing in a raft of rules, trainings and policies that slow things down so that a situation doesn’t escalate. This can include stuff like the foot pursuit policy, but also telling officers to calmly talk to a suspect[101] or not to use force at all if someone isn’t posing a physical threat[102].

And while it’s early days here, police reports from this year suggest this approach is working in San Francisco[103][104] , Seattle, [105] And New Orleans too. In fact, in New Orleans - their department credits de-escalation training for[106] lowering police shootings since 2012.[107] 

Now, some officers that we spoke to were concerned[108] though - that not being allowed to use as much force - wouldn’t let them do their job properly. They’d be put at risk. And Crime would go up. Is that true?  Well it’s very rare for an officer to die on the job:[109]... You’re more likely to die as a roofer or a fisherman.[110] [111] [112] But Injuries are pretty common, so will that number go up? We’re actually not sure. In one city, there was up-tick in police hospitalizations[113] after these policies kicked in, but it could have been because the department changed how those cases were reported. In other cities[114] - it didn’t affect officer injuries. And we do know that similar changes in the past might have made things safer for cops: in the 1970s, 50 cities all across the US put in tighter rules around when cops could shoot people.[115] And this was linked to a huge drop in shootings of both civilians and cops.[116] And Phil says this makes sense[117]. De-escalating situations can work for everyone.

PG Think about it, if we're having a disagreement, and you think you can take me, I might get hurt. And it might be me. But if no one has to fight, everyone goes home happy.

And what about crime? Are these policies affecting the crime rate? Well crime goes up and down for all sorts of reasons - that science doesn’t fully understand. Seriously. It’s kind of a mystery[118][119]. So far though, these policies aren’t affecting crime in any consistent way.[120] So all in all, Phil, is pretty optimistic… 

PG People are surprised - I don’t sound like an optimistic guy in general, but I’m incredibly optimistic about where we are with policing. I think it’s one of the best race stories in the country. I don’t believe arc of the moral universe is just born that way I think we need to bend it ourselves. But I think we’re going to win. Right. I think we’re going to win.

So when it comes to police shootings… what do we know?

Police in the US do treat black people differently from white people - while we don't have lots of good data on police shootings --- these differences are really stark when you look at data on stops and searches. So far - implicit bias training and police body cams - aren’t really the silver bullet that many had hoped for. What’s emerging as a breakthrough here - something that could really help - is making clear rules and policies that tell cops when they should - and shouldn’t use force. And encouraging them to de-escalate a situation. For now, this idea seems to be helping. 

That’s Science Vs.



This episode was produced by Meryl Horn with help from me Wendy Zukerman, along with Rose Rimler, Michelle Dang, Lexi Krupp, and Kaitlyn Sawrey. We’re edited by Caitlin Kenney and Blythe Terrell. Fact checking by Diane Kelly. Mix and sound design by Peter Leonard with help from Cedric Wilson. Music written by Peter Leonard, Benny Reid, Emma Munger, and Bobby Lord. A big thanks to Professor Laurence Sherman, Dr. Joe Cesario, Dr. Sam Walker, Chuck Wexler, Dr. Peter Moskos, Dennis Flores, Hawk Newsome, Professor William Terrill, Dr. Arne Nieuwenhuys, Professor Franklin Zimring, Dr. Joan Vickers, and Dr. Justin Nix. Thanks to all police officers we spoke to- we really appreciate your help. And special thanks to Amber Davis, Chuma Ossé, Daniel Domke, Christina Djossa, the Zukerman family and Joseph Lavelle Wilson.

I’m Wendy Zukerman, fact you next time.

[1] Timeline at https://apnews.com/9aa32033692547699a3b61da8fd1fc62 

[2] https://blacklivesmatter.com/herstory/

[3] 67% of officers say these encounters are “isolated incidents” https://www.pewsocialtrends.org/2017/01/11/behind-the-badge/

[4]“Unfortunately, the federal government has failed to maintain a reliable, comprehensive database of OIS, in spite of social scientists imploring them to compile better criminal justice data for decades (Alpert 1948; Alpert 2015; Fyfe 2002).  http://sci-hub.tw/10.1080/0735648X.2018.1547269 

[5] January 1, 2019: The data collection launched nationwide. All law enforcement agencies are encouraged to participate… The FBI has no legal authority to mandate reporting of any data to the UCR Program

[6] https://www.theguardian.com/us-news/ng-interactive/2015/jun/01/the-counted-map-us-police-killings 

[7]https://www.theguardian.com/us-news/ng-interactive/2015/jun/01/the-counted-police-killings-us-database# 2016

[8] https://www.washingtonpost.com/graphics/2018/national/police-shootings-2018/

[9] from WaPo: 41%-46% no gun depending on year, from Guardian 52%-54% didn’t have a gun (and this includes not just shot, but killed by police)

[10] https://www.washingtonpost.com/graphics/2018/national/police-shootings-2018/ 

[11] When Police Kill. Franklin Zimring: p 46 based on Guardian data “The death rate for blacks/african americans per population is 2.3 times the white non-hispanic rate”

[12] http://sci-hub.tw/10.1080/0735648X.2018.1547269 based on WaPo data- table 1: Benchmarked to population: 2.77-3.04

[13] http://sci-hub.tw/10.1111/1745-9133.12174 St Louis study-  “These data indicate that police use of deadly force is higher in areas with larger minority populations and with higher levels of violent crime, as well as in areas with higher levels of economic disadvantage.” - see figure 1 for overlap of firearm violence rate and percent black - also dots for officer involved shootings

[14] http://sci-hub.tw/https://doi.org/10.1111/j.0038-4941.2005.00353.x  Wilson suggested that neighborhoods with very high levels of poverty are often exponentially worse in terms of social disorganization and social pathology (including crime) than neighborhoods with moderate levels of poverty.; Moreover, the disadvantage-homicide relationship appears especially strong in extremely poor areas (and in predominately African-American neighborhoods). (However) Specifically, Krivo and Peterson (2000) reported that the relationship between disadvantage and homicide for African Americans was nonlinear, with the effect of disadvantage on homicide leveling off for extremely disadvantaged areas.

[15] https://link.springer.com/article/10.1007/s11524-008-9302-y study of 10 US cities: “Homicide rates were greater in neighborhoods with higher concentrations of African-Americans and Hispanics than in other groups”

[16] Regarding traffic stops: These disparities stem, in part, from a strategy that concentrates traffic stops in high-crime areas.; poverty also makes minor violations that produce traffic stops (expired tags, broken lights) more likely

[17] Police are more likely to interact with individuals who are on the ‘social margin’, such as members of lower socioeconomic classes and members of minority groups (Greenfeld et al. 1997, Worden and Catlin 2002). As summarised by Walker et al. (1996) the disparate assignment of police in minority and/or lower income areas occurs for a number of reasons. Officers are assigned to neighbourhoods based on the number of reported crimes and calls for service that occur in a particular area

[18] https://pdfs.semanticscholar.org/a780/f37f5eaaf9d72c71907d6ea29f17f8774520.pdf . The

authors found police officers are significantly more likely to use higher levels of force when suspects are encountered in disadvantaged neighborhoods and those with higher homicide rates, net of situational factors (e.g., suspect resistance) and officer based determinants (e.g., age, education, and training). Also found is that the effect of the suspect’s race is mediated by neighborhood context.

[19] http://sci-hub.tw/10.1146/annurev-criminol-032317-092409  ... in most analyses across communities, homicide rates are positively (and strongly) correlated with fatal police shootings (Klinger et al. 2015, Liska & Yu 1992, Sherman & Langworthy 1979).

[20] https://web.stanford.edu/~eberhard/about-jennifer-eberhardt.html

[21] Brought in as a subject matter expert to analyze data on police stops by race: https://books.google.com/books?id=vpdeDwAAQBAJ&

[22] http://sci-hub.tw/10.1177/0963721418763931 and original paper:: https://stanford.app.box.com/v/Data-for-Change OPD officers stopped, searched, handcuffed, and arrested more African Americans than Whites, a finding that remained significant even after we controlled for neighborhood crime rates and demographics; for stops “This suggests that something other than greater police presence and the resulting higher number of stops in high-crime-rate neighborhoods is driving the high numbers of African American stops. The oft-repeated explanation that police simply target high-crime areas, and those areas happen to be predominantly African American, does not seem to be enough to explain the prevalence of African American stops in its entirety." for handcuffing: African American men were handcuffed in 1 out of every 4 stops vs. 1 in every 15 stops for White men.. And for searches: Black men were searched in 1 out of 5 stops, vs. 1 out of 20 stops for White men and “we still see no significant differences in search recovery rates by race” .

[23] Statistical analyses of “stop data” from 28,119 forms that 510 OPD officers filed after stopping drivers and pedestrians in Oakland, Calif., between April 1, 2013,and April 30, 2014 (for a summary, see Chapter 1of Strategies for Change; for the technical report, see Data for Change);

[24] Controlling for population Also yes see p 49-51 https://stanford.app.box.com/v/Data-for-Change

[25] White males were searched in 5.5% of stops, compared to 20.4% for African American males and 14.4% for Hispanic males. https://stanford.app.box.com/v/Data-for-Change 

[26] New York City https://5harad.com/papers/frisky.pdf   We further find that blacks and Hispanics were disproportionately stopped in these low hit rate contexts, a phenomenon that we trace to two factors: (1) lower thresholds for stopping individuals—regardless of race—in high-crime, predominately minority areas, particularly public housing; and (2) lower thresholds for stopping minorities relative to similarly situated whites… Moreover, while 49% of blacks stopped under suspicion of CPW have less than a 1% chance of in fact possessing a weapon, the corresponding fraction for Hispanics is 34%, and is just 19% for stopped whites.

[27] https://5harad.com/papers/100M-stops.pdf Specifically, in the 16 state patrol agencies for which we have the necessary data, search rates were 3.8%, 3.6%, and 1.6% for stopped black, Hispanic, and white drivers, respectively.

[28] National study: https://5harad.com/papers/100M-stops.pdf we find evidence of bias against black drivers both in highway patrol and in municipal police stops. . Examining both the rate at which drivers are searched and the likelihood that searches turn up contraband, we find evidence that the bar for searching black and Hispanic drivers is lower than for searching whites.- data was from 2011-2017, study published in 2019

[29] Counted the number of unique states in table 1

[30] https://sfdistrictattorney.org/sites/default/files/Document/BRP_report.pdf SF-  Black and Hispanic people are more likely to be searched without consent than any other group, and, of those searched, Black and Hispanic people had the lowest “hit rates” (i.e., the rate at which searches found contraband). Graph: p 30. Study from 2016- data was from many different dates…

[31] New York City https://www.nyclu.org/sites/default/files/publications/2012_Report_NYCLU_0.pdf  Though they account for only 4.7 percent of the city’s population, black and Latino males between the ages of 14 and 24 accounted for 40.6 percent of stops in 2012. …  a weapon was found in only 1.8 percent of blacks and Latinos frisked, as compared to a weapon being found in 3.9 percent of whites frisked. Study from 2012

[32] https://policylab.stanford.edu/media/nashville-traffic-stops.pdf 2017: Nashville. Disparity in stops for non-moving violations- partly went away after controlling for crime (went from 68% more likely to 37%).  https://policylab.stanford.edu/media/nashville-traffic-stops.pdf  Black drivers were stopped 44% more often per driving-age resident when compared to white drivers; this gap is particularly pronounced among stops for non-moving violations (68%), such as broken tail lights and expired registration tags… These disparities stem, in part, from a strategy that concentrates traffic stops in high-crime areas… (the found the disparity fell to 37% once they accounted for neighborhood crime, but was still there)

[33]https://www.aclum.org/sites/default/files/wp-content/uploads/2015/06/reports-black-brown-and-targeted.pdf While the 2010 census reveals that Blacks made up 24.4% of Boston’s population, they comprised 63.3% of police-civilian encounters from 2007 to 2010…. The research team conducted several analyses to measure the effect of race on these encounters. Their preliminary findings confirm that Blacks were more likely to experience both stops and searches—even after controlling for non-race factors such as neighborhood crime rates or the past arrest records and alleged gang affiliation of the civilians subjected to police encounters.  

[34]https://www.aclusocal.org/sites/default/files/wp-content/uploads/2015/09/11837125-LAPD-Racial-Profiling-Report-ACLU.pdf After controlling for violent and property crime rates in specific LAPD reporting districts, as well as a range of other variables, we find That:  “Per 10,000 residents, the black stop rate is 3,400 stops higher than the white stop rate, and the Hispanic stop rate is almost 360 stops higher.” disparities are also listed for arrests, searches; “It is implausible that the higher frisk and search rates were justified by higher minority criminality, when these frisks and searches were less likely to uncover weapons, drugs or other types of contraband.”

[35] https://www.aclu.org/other/oakland-police-department-body-camera-policy 

[36] Our dataset consists of transcribed body camera footage from vehicle stops of white and black community members conducted by the Oakland Police Department during the month of April

2014. We examined 981 stops of black (N = 682) and white (N = 299) drivers from this period, 68.1% of the 1,440 stops of white and black drivers in this period. https://www.pnas.org/content/pnas/114/25/6521.full.pdf

[37]Discussed at conference https://aaas.confex.com/aaas/2019/meetingapp.cgi/Paper/23960 

[38]We observe racial disparities in officer respect even in police utterances from the initial 5% of an interaction, suggesting that officers speak differently to community members of different races even before the driver has had the opportunity to say much at all” https://www.pnas.org/content/pnas/114/25/6521.full.pdf; model discussed in 3.7 Linguistic Classification Accuracy of Race https://www.pnas.org/content/pnas/suppl/2017/05/30/1702413114.DCSupplemental/pnas.1702413114.sapp.pdf

[39]Figure 2:  https://www.pnas.org/content/pnas/114/25/6521.full.pdf 

[40] https://www.pnas.org/content/pnas/114/25/6521.full.pdf  Are the racial disparities in the respectfulness of officer speech we observe driven by a small number of officers? We calculated the officer-level difference between white and black stops for every officer (N = 90) in the dataset who had interactions with both blacks and whites (Fig. 4). We find a roughly normal distribution of these deltas for officers of all races.

[41]These disparities are part of a constella- tion of differences in officer language spoken toward black versus white community members; a simple classifier trained on only the words used by officers is able to correctly predict the race of the community member in over two thirds of the interactions“: https://www.pnas.org/content/pnas/114/25/6521.full.pdf 

[42] Growth since 2010

[43] In other words, differences reflect passively acquired attitudes, not personal feelings.  

[44] Study of 80 officers:presented them with pictures of either neutral objects or weapons, or black or white faces  This indicates that the officers tended to have moderate (35%) to strong (37%) bias associating Black Americans with weapons. Approximately 12% of officers had slight anti-Black bias and a further 12% had no bias. Finally, a combined 3% of the sample had anti-White bias (associating White Americans with weapons)... ‘‘slight black bias’’  12.0%, ‘‘moderate black bias’’  34.8%, and ‘‘strong black bias’’  37.3%). = 84%

[45] https://web.stanford.edu/~eberhard/downloads/2004-SeeingBlackRaceCrimeandVisualProcessing.pdf 

[46] http://sci-hub.tw/10.1037//0022-3514.34.4.590 (1976) students were asked to rate the behavior of people in a video where one was shoving the other. “Examining the frequencies for the black-protagonist/ white-victim conditions, 75% of the subjects chose the violent behavior major category, whereas when the protagonist was white and the victim was black, the behavior was labeled violent by 17% of the subjects”

[47]  https://www.apa.org/pubs/journals/releases/psp-a0035663.pdf Participants overestimated the age of Black targets and deemed Black targets more culpable for their actions than White or Latino targets, particularly when those targets were accused of serious crimes.

[48] Despite this, implicit bias training is growing in popularity; becoming a staple in academies and departments across the United States

[49] https://www.cbsnews.com/news/racial-bias-training-de-escalation-training-policing-in-america/

[50] https://shsu-ir.tdl.org/bitstream/handle/20.500.11875/2497/1808.pdf?sequence=1

[51] https://www1.nyc.gov/site/nypd/news/s0626/nypd-response-oig-report



[54] https://senate.uoregon.edu/tag/implicit-bias/


[56] More info DOJ implicit bias training: https://www.justice.gov/opa/file/871121/download

[57] Price depends on the size of the department. For some larger cities it can be hundreds of thousands of dollars See for example this news story: Phoenix police spending $150,000 per year for several years (totaling $450,000) https://www.azcentral.com/story/news/local/phoenix/2018/07/05/phoenix-police-get-450-k-implicit-bias-training/755596002/ 

[58] $4.5 million for the NYPD Police department: https://www1.nyc.gov/office-of-the-mayor/news/360-18/transcript-mayor-de-blasio-appears-live-inside-city-hall

[59] Also see the pricing sheet for NTRE here: 1 full day for 100-200 people = $10K https://d1fa577f-c8d0-450d-992f-00340043ce61.filesusr.com/ugd/647e86_64afcdb78cd34e76ac237d803e09d530.pdf 

[60] https://policingequity.org/about/executive-team/phillip-atiba-goff-phd

[61] Story told on This American Life: https://www.thisamericanlife.org/548/transcript 

[62] UK Equality and Human Rights (2018) :The evidence for Unconscious Bias Training changing short-term implicit bias is consistent. Implicit bias is likely to be reduced, but not eradicated, through UBT.

[63] Small (n = 91) 2012 study: The intervention is based on the premise that implicit bias is like a habit that can be reduced through a combination of awareness of implicit bias, concern about the effects of that bias, and the application of strategies to reduce bias; These findings offered initial evidence that bias-reduction training could lead to change at least two months later.

[64] (2017) (n=292)  Effect generally declined after 2 weeks, but Intervention participants were more likely to interact with Black strangers, were more likely to report noticing bias and to label it as wrong, and, two years later, were more likely to confront bias in others.

[65] (2016) (n=around 6000) All 9 methods reduced implicit racial bias. However, none were effective after a delay of several hours to several days. A reanalysis of this study by another group in 2019 found: “"Although social environments are stable, individual implicit biases are ephemeral. … Our core claim in this article is that individual level implicit bias is not a rigid attitude"”

[66] Forscher, P. S., Lai, C. K., Axt, J., Ebersole, C. R., Herman, M., Devine, P. G., & Nosek, B. A. (2019). A Meta-Analysis of Procedures to Change Implicit Measures. https://doi.org/10.1037/pspa0000160 Procedures that associate sets of concepts, invoke goals or motivations, or tax mental resources changed implicit measures the most, whereas procedures that induced threat, affirmation, or specific moods/emotions changed implicit measures the least. Tweet by author C. Lai explaining results: https://twitter.com/calvinklai/status/1127958418113875969?lang=en

[67] https://www.usccr.gov/pubs/2018/11-15-Police-Force.pdf Dave Wilson, University of Delaware: You can’t train away implicit biases, they are implicit. They are there.

[68] https://www.nytimes.com/2019/11/20/style/diversity-consultants.html “For now, industry leaders have a range of backgrounds: social justice, sales, marketing, political activism, minority studies, writing and more.”  Sometimes internal employees trained to teach: Examples of Train-the-Trainer programs

[69] Some editorials from our experts:

Jennifer’s LATimes article: “But the very path to making implicit bias non-threatening may also normalize it in a way that makes it harder to eradicate.Research shows that when something is regarded as normal, people cease to judge it harshly. So when bias is presented as an ordinary part of human functioning, that can lead us to care less about the dangers it presents and feel less motivated to do the work that clearing out our preconceptions require.”

Editorial by Lois and Renee Mitchell, JD, PHD “A concerning possibility that has been raised by social justice scholars is that implicit bias training may have an unintended negative effect by increasing the expression of bias. Researcher Joshua Correll has suggested “there are a number of very compelling studies that show that if you just ask somebody to try really hard not to show racial bias you can actually inadvertently increase racial bias.””    

[70] Pg 21 Hausmann et al.’s (2014) Second, after completing the training program, members of the training PACTs reported increased agreement with the belief that biases cannot be changed … It was not intended that the training would cultivate the perception that biases cannot be changed; however, the training’s focus on the unconscious nature of biases may have resulted in trainees believing that biases cannot be changed since unconscious processes are more challenging to recognize and to change than are conscious beliefs and actions.

[71] http://sci-hub.tw/10.1037/a0037908 The present research demonstrates that individuals who received a high prevalence of stereotyping message expressed more stereotypes than those who received a low prevalence of stereotyping message

[72] “Stereotype Rebound” effect  (1994 study) The results provided strong support for the existence of this effect. Relative to control subjects (i.e., stereotype users), stereotype suppressors responded more pejoratively to a stereotyped target on a range of dependent measures.

[73] As of May 18, 2015, ten states (Arizona, Colorado, Florida, Maryland, New Hampshire, North Dakota, Oklahoma, Pennsylvania, Utah, and Vermont,) enacted laws related to BWCs, several of which directly address public disclosure and privacy issues (National Conference of State Legislature (NCSL),

[74] 33 states plus DC have laws or are developing legistation related to BWCs as of 2019 (National Conference of State Legislature (NCSL)

[75] Bureau of Justice Statistics reports that as of 2016, 60% of local police departments and 49% of sheriffs' offices had fully deployed their BWCs (Hyland, 2018).https://www.bjs.gov/content/pub/pdf/bwclea16.pdf 

[76] https://bwc.thelab.dc.gov/TheLabDC_MPD_BWC_Working_Paper_10.20.17.pdf With 2,224 MPD members participating in the study, this is one of the largest randomized evaluations of BWCs conducted to date.

[77] https://thepolicylab.brown.edu/team/david-yokum/

[78]  He followed each district for about 7 months. "The last district to receive cameras was 2D, which deployed BWCs on May 17, 2016. This gives us a window of 212 days (from May 17-December 15, 2016), or approximately a seven-month study duration for each district. Thus, we effectively measure outcomes for the first seven months in each district, and report the results averaged across all seven police districts. See Supplementary Materials for the start date of BWC deployment in each district." p 8 https://bwc.thelab.dc.gov/TheLabDC_MPD_BWC_Working_Paper_10.20.17.pdf

[79] Metropolitan Police Department. General Order - Use of Force. GO-RAR-901.07. https://go.mpdconline.com/ GO/GO_901_07.pdf. 1 Dec 2016. Per this policy, “The following actions are designated “reportable uses of force”: (1) Deadly force; (2) Serious use of force; (3) Use of a less-than-lethal weapon; (4) Any use of force indicating potential criminal conduct by a member; and (5) Any use of force resulting in injury or a complaint of injury or pain where the injury or pain is directly associated with a member’s use of force. The following actions are designated “reportable force incidents” as long as the use of force does not result in injury or a complaint of injury or pain: (1) All solo or team takedowns, where there is no complaint of pain or injury; and (2) The drawing and pointing of a firearm at, or in the direction of, another person when no other force was used. Minor injury or discomfort resulting from the application and general wearing of handcuffs is not, in and of itself, considered a “reportable use of force” or a “reportable force incident”

[80]  https://journals.sagepub.com/doi/full/10.1177/1477370816643734 Averaged over 10 trials, BWVs [Body-worn videos] had no effect on police use of force (d = 0.021; SE = 0.056; 95% CI: –0.089–0.130)

[81] Approx. number of US citizens having police interactions during 2015 = 53.9 million. That means a police-civilian interaction resulting in a death is super-rare. Data here: https://www.bjs.gov/content/pub/pdf/cpp15.pdf 

[82] https://www.motherjones.com/politics/2015/05/police-shootings-caught-on-tape-video/

[83] https://nursing.wsu.edu/people/lois-james

[84] Zimring When Police Kill, “The largest number of lethal events was produced by disturbance calls” - pg 51, pg 52 for graph. “Domestic Disturbances” are 12.4%- one of the highest categories. Based on Guardian data.




[88]https://www.researchgate.net/profile/Peter_Jensen18/publication/268277336_Performance_Under_Acute_Stress_A_Qualitative_Study_of_Soldiers'_Experiences_of_Hand-to-Hand_Combat/links/5470b0050cf2d67fc031cee4/Performance-Under-Acute-Stress-A-Qualitative-Study-of-Soldiers-Experiences-of-Hand-to-Hand-Combat.pdf The present findings also suggest that the speed and brevity of hand-to- hand fighting can exacerbate soldiers’ stress


[90]https://www.fletc.gov/sites/default/files/imported_files/reference/research-papers/survival_scores_research.pdf  This study from 2004, on federal law enforcement students, 100 participants, escalated scenarios gradually to produce more and more stress (pg 22). Heart rate (pg 33), Blood pressure (pg 34) and cortisol all went up (pg 35), and were elevated during the shoot out (When the students were most likely to fail).”Overall, performance, deteriorated, as expected, and only 28.2% performed well enough to pass this event..”

[91] https://pdfs.semanticscholar.org/2774/66f8e47d3cede0cec5f5950a7758226f0519.pdf  “That is, whereas moderate stress levels provoke moderate cortisol reactivity and usually result in positive effects of improved memory (Wolf, 2009), stress research during strenuous military training has found both robust cortisol increases and significant decreases in cognitive performances”; Fig 4: more stressful training scenario associated with lower scores for working memory.  

[92] http://sci-hub.tw/10.1037/a0025699 When performing under anxiety, police officers showed a response bias toward shooting, implying that they accidentally shot more often at suspects that surrendered.

[93]  Charlotte-Mecklenburg officers are instructed to refrain from using a TASER® except when dealing with physically aggressive suspects, while both Colorado Springs and Albuquerque are permitted to use a TASER®  on suspects displaying defensive resistance;  With Charlotte-Mecklenburg serving as the reference category, we see that officers in both Colorado Springs and Albuquerque are more likely to use a TASER® when dealing with defensively resistant suspects. In fact, we see that officers in Albuquerque are nearly six times more likely to use a TASER®, while Colorado Springs officers are nearly 23 times more likely to use a TASER®. Dependent Variables: soft hand tactics, hard hand tactics, chemical sprays, conducted energy devices, and other impact weapons (batons, flashlights, and bean bags).

[94] https://www.ire.org/resource-center/stories/25264/ By any measure, Nevada's largest law enforcement agency uses deadly force more often than counterparts in the region and in other major cities surveyed. [from 2011]

[95]https://www.reviewjournal.com/opinion/editorials/deadly-force-las-vegas-metro-looks-at-revamping-some-policies/ Following December’s high-profile police killing of unarmed Iraq War veteran Stanley Gibson — which followed shortly after the 2011 publication of the Review-Journal’s multi-part examination of officers’ use of deadly force

[96] https://qz.com/565011/how-one-of-the-largest-police-forces-in-america-stopped-shooting-people/ Cole was killed even though he was unarmed. It would be a record year for officer-involved shootings in Las Vegas, and Cole’s death was one of several high-profile incidents which led to calls to reform what seemed to be a trigger-happy police force.

[97] https://behavioralpolicy.org/wp-content/uploads/2017/05/Goff-web.pdf the bulk of foot pursuits stop when the suspect realizes he or she is surrounded and gives up.

[98] https://behavioralpolicy.org/wp-content/uploads/2017/05/Goff-web.pdf The policy went into place at the end of 2011. Figure 2 reveals that the LVMPD experienced a 23% drop in use-of-force incidents and a further decline the following year.

[99]For example, in this five year period the reported use of oleoresin capsicum (OC) spray declined by 59 percent, electronic control devices (ECD) by 50 percent, and handcuffs by 42 percent. (Use of canines also increased)

[100] https://www.theiacp.org/sites/default/files/IACP_UC_De-escalation%20Systematic%20Review.pdf “The National Consensus Policy specifically defines police de-escalation as “taking action or communicating verbally or non-verbally during a potential force encounter in an attempt to stabilize the situation and reduce the immediacy of the threat so that more time, options, and resources can be called upon to resolve the situation without the use of force or with a reduction in the force necessary”

[101] The term de-escalation generally refers to the act of moving from a state of high tension to a state of reduced tension (Richards, 2007). In law enforcement, minimizing danger and tension in potentially volatile situations is a daily responsibility;  Removing any distractions or disruptive persons from the area so that the officer can maintain focus on the individual experiencing the crisis is an important component of de-escalation. The officer should remain calm and speak slowly, in short sentences, to encour- age communication. The responding officer should also present a genuine willingness to understand and help.

[102] When safe and feasible under the totality of the circumstances, officers shall attempt to slow down or stabilize the situation so that more time, options and resources are available for incident resolution…. Using verbal techniques, such as Listen and Explain with Equity and Dignity (LEED) to calm an agitated subject and promote rational decision making - Avoiding language, such as taunting or insults, that could escalate the incident

[103] Chief William Scott described the changes "“We know that our focus on de-escalation and proportionality is creating more positive outcomes,” https://www.sanfranciscopolice.org/news/san-francisco-police-department-releases-3rd-quarter-2019  and https://www.sfchronicle.com/crime/article/San-Francisco-police-tout-use-of-force-drop-13808787.php  - “, “[Chief Bill Scott] credits the dip to a new and stricter use-of-force policy and training that emphasizes de-escalating tense situations, as well as broader crisis intervention training that all officers have received.”

[104] Total use of force: page 5, has been going down https://www.sanfranciscopolice.org/sites/default/files/2019-11/Sfpd96AExecSumQ32019.PDF "Overall uses of force have dropped 47 percent since such data began being reported in July 2016 as required under Chapter 96A." https://www.sanfranciscopolice.org/news/san-francisco-police-department-releases-3rd-quarter-2019

[105] https://www.seattle.gov/Documents/Departments/Police/Publications/2019_Annual_UoF_Report.pdf   2018: the number of reported Type I (low level) uses of force increased by 43%, while reported uses of Type II and III force decreased.

[106] "NOPD agrees to ensure that officers use non-force techniques to affect compliance with police orders whenever feasible; use force only when necessary, and in a manner that avoids unnecessary injury to officers and civilians; and de-escalate the use of force at the earliest possible moment. To achieve these outcomes, NOPD agrees to implement the requirements set out below.” http://nopdconsent.azurewebsites.net/Media/Default/Documents/Docket%20Items/575%20Monitor's%20Report%20On%20NOPD%20Progress%20Under%20The%20Consent%20Decree.pdf 

[107] Table on Page 15: officer involved shootings of persons down each year from 2012 to 2018. page 18: canine bites, doen also. Weird - they also keep track of shooting animals - slightly down (page 15) http://nopdconsent.azurewebsites.net/Media/Default/Documents/Docket%20Items/575%20Monitor's%20Report%20On%20NOPD%20Progress%20Under%20The%20Consent%20Decree.pdf 

[108] A narrow majority of officers (56%) feel that in some neighborhoods being aggressive is more effective than being courteous, while 44% agree or strongly agree that hard, physical tactics are necessary to deal with some people.

[109] Accidentally killed and feloniously killed numbers here: https://ucr.fbi.gov/leoka/2018/topic-pages/officers-accidentally-killed 

[110] The rate of fatal work injuries for police officers in 2014 was 13.5 per 100,000 full-time equivalent workers- varied between 10.6-20 for 2006-2014

[111]  https://www.bls.gov/news.release/archives/cfoi_09172015.pdf 2014 Fisherman 80.8 per 100,000 and Roofers 46.2 per 100,000

[112] All in 2014 stats: https://www.bls.gov/news.release/archives/cfoi_09172015.pdf for 2014: Fishers 80.8 per 100,000 and Roofers 46.2 per 100,000

[113]  this report says injuries stayed the same but hospitalizations increased https://cops.usdoj.gov/RIC/Publications/cops-w0834-pub.pdf  page 39 -  it is possible that there may have been a change with regard to how officer injury was recorded; for example, in 2006 the Department shifted from allowing officers to record a number of treatment options (e.g. first aid, treated and released, hospitalized) to only including treated in hospital.

[114] NEW ORLEANS- From their media department, numbers from 2012-2018 for officer injuries from assaults :47, 52, 38, 33        ,32, 40, 45. SEATTLE- Officer injuries are flat to slightly down over the study period, although the decrease is not statistically significant, based on SPD injury and hospitalization data. SAN FRANCISCO- from 2016,2017 , 2018, 2019 third quarter data: 111, 61, 76, 81 >> overall, does not appear to be increased

[115] From 1970 through 1985, 50 cities of more than 250,000 residents each took

actions that cut in half the annual total count of citizens killed by police in those cities from 353

to 172 per year (Sherman & Cohn 1986). The major change during this period was a growing ban

on shooting nonviolent fleeing suspects. http://sci-hub.tw/10.1146/annurev-criminol-032317-092409

[116] This now-forgotten change in police-citizen violence saw both killings by and killings of police fall dramatically, by 51% and 65%, respectively. http://sci-hub.tw/10.1146/annurev-criminol-032317-092409 

[117]Also true in the other direction - On the other hand: using nationwide panel data on local police militarization, I demonstrate that militarized policing fails to enhance officer safety or reduce local crime

[118] After nearly 50 years of theoretical and empirical work,...huge questions remain unanswered. It remains a mystery, for instance, why crime rose so much in the 1960s.

[119] It is an embarrassment to criminology that the causes of the great American crime

drop are so poorly understood.

[120]New Orleans:  it's been going up a little from 2012-2019 Las Vegas: seems to be constant from 2011-2012, Seattle “crime, by most measures, has not increased. San Francisco: Property crime has been going up in general, violent crime has remained constant from 2016-2019