How Many Homicides were there in 2010?

How many homicides were there in 2010 in the United States? Well, that’s easy. Let’s just do some Googling:

  1. 12,966, FBI, Crime in the United States 2010.
  2. 13,164, FBI, Crime in the United States 2011 (2010 figure).
  3. 14,720, Bureau of Justice Statistics (Table 1, based on FBI, Supplementary Homicide Statistics).
  4. 16,259, CDC (based on death certificates in the National Vital Statistics System).

Between the smallest and largest figures there is a difference of 3,292 deaths or 25%!

The differences are striking but not entirely arbitrary or without explanation. I assume the second figure adds late additions to the 2010 data and so should be considered more authoritative but that is a relatively small difference.

The difference between 2 and 3 is puzzling and seems to be that the number in 2 is drawn from the Supplementary Homicide Report (SHR) statistics on victims while the larger figure is drawn from homicide reports in the UCR. Not all agencies collect the more detailed statistics in the SHR while the UCR is nearly complete. Thus the victim figure is smaller than the report figure (this doesn’t appear to conform exactly to where the data is supposed to be sourced but it’s what the FBI tells me). It’s unclear why the FBI would report both figures when they know one is misleading.

The difference between 3 and 4 comes from different definitions of homicide. The FBI collects data on crimes. If a killing is ruled justified, i.e. not a crime, it doesn’t go into the FBI homicide statistics. The CDC collects data from death certificates which list as homicide any death caused by “injuries inflicted by another person with intent to injure or kill, by any means.” Thus, the CDC data includes justifiable homicide. In 2010 according to the FBI there were 387 justifiable homicides by law enforcement and 278 by private citizen for a total of 665 justifiable homicides, so that accounts for some but not all of the difference.

(By the way, the 278 justifiable homicides in 2010 by private citizens compared to 387 by law enforcement and 14,720 unjustifiable homicides would seem to be an important context for many claims about stand your ground laws. N.b. this doesn’t mean that the laws couldn’t be associated with more unjustifiable homicides).

The FBI (3) and NVSS (4) figures track each other closely over time but its important to be aware of the differences and to be consistent in one’s calculations.


I have a pretty clear recollection of reading an article on the difference between FBI and CDC homicide statistics years ago. The reasearcher had done qualitative interviews with doctors, some of whom told him that if a death looked violent, they classified it as a homicide, leaving it to the police to decide whether it was a homicide or a suicide. So, the CDC figure might include quite a few suicides.

I tried to find this article later, but couldn't. If anyone has a link, this would be much appreciated.

The problem is widespread. Years ago I need figures for deaths in road accidents in the UK. It turned out that the two different sources - the police and the NHS - gave discernibly different numbers.

Note that you *can* find #3 in the 2010 report (""An estimated 14,748 persons were murdered nationwide in 2010""):

Data collection footnotes are important: "Law enforcement agencies are asked (but not required) to provide complete supplementary homicide data for each murder they report to the UCR Program." Says so right there on the front page of the report.

The victim breakdown in table 1 and beyond is based on more detailed reports, not all of which are supplied.

It is extremely important to note that FBI UCR reporting is *voluntary.* CDC reporting is not voluntary.

In other words, hospitals have to put something on the death certificate, while small police departments need not submit detailed UCR reports (or, any at all!). Lots of smaller police departments do not submit UCR reports at all.

A time series of data I've tracked since 1905 has the two differing by an average of 9% - FBI UCR will be more accurate for urban areas with large police departments and resources, inaccurate for smaller rural states or areas (eg Montana and North Dakota).

Thank you.

to reiterate, you have to be careful which table you are looking at in the reports:
^ this table has total offenses and is not based on the "Expanded Homicide Data"

This could have something to do with it: "The Detroit Police Department is systematically undercounting homicides, leading to a falsely low murder rate in a city that regularly ranks among the nation's deadliest, a Detroit News review of police and medical examiner records shows."

If other departments are doing this, the FBI stats would be off.

Baltimore was accused of not reporting rapes to drive down the stats. IIRC, they would classify them as simple assaults. The incidence of "unfounded rape claims" was something like 5 times higher than the national average. Murder is a little tougher to hide. "Unfounded cases of murder" would be an interesting file to read. But, the incentive to fudge the numbers is overwhelming so they will find a way. Therefore, one should view crime stats as "best case scenario."


Link hopefully in one piece:

And another:

What year does it count if I kill someone and it isn't discovered for 5 years?

Another minor methodological conundrum is that, say, Assault with a Deadly Weapon can turn into Homicide if the victim later dies.

For example, Col. Joshua Lawrence Chamberlain, a hero of Gettysburg, is often said to have died from his Civil War wounds in about 1913.

Obviously, that's a pretty subjective judgment, the kind that is only made about celebrities, but you can see the issues involved.

this is another good example of the core problem with routine statistical reports and reporting; people tend to blindly believe numbers. Disconnects and errors in federal government numbers are especially egregious because so many people falsely assume that numbers issued with federal imprimatur are absolute facts.

/ on homicide data in particular, what is the practical value of knowing such aggregate numbers? If you knew precisely the # of U.S. homicides every year or month or hour -- what difference would it make? What would anyone do with such data, of real consequence?

So, that the reason you cited the correct homicide count prepared by the for profit private sector gun and weapon industry which wants an objective body count?

Knowing the precise numbers isn't all that critical for most applications, but:

a. Using the same set of numbers, collected in the same way, for comparisons and in calculations is really important, or you're liable to get nonsense out.

b. Knowing how much play there is in the official numbers (how much is really rock solid, how much is trying to fill in missing data in a sensible way or educated guesswork) is useful, if you're trying to notice a fairly small effect. A study showing that some new law or policy led to a change in the murder rate less than the amount of play in the numbers by slight changes in your counting or data collection rules seems rather likely not to be measuring what you intend it to measure.

Lots of large police departments under-report crime. see for example here:

It's hard to hide bodies though, and reclassify homicides. Justifiable homicides are reported separately to the FBI, so I would not necessarily call those inaccuracies. As the article notes, in Detroit there are maybe a handful that the ME reported as homicide where the police reported as suicide or "inconclusive" - the real inaccuracies. There are lots of tin foil hat people who think that the NYPD for example under Bloomberg also has been systematically underreporting homicides.

If you wanted to actually check this, its pretty easy to do: CDC has a tool where one can retrieve homicide data by county (you'll see in select location that you can open it to the county level)
New York City consists of five boroughs: Bronx County; Kings County; New York County; Queens County; Richmond County, NY;

CDC wonder says the total for 2010 was 544 for the 5-county borough.

The FBI UCR says 536 ( allows local agency crime search). I don't know offhand how many justifiable homicides there were in NYC in 2010 offhand, a difference of 8 seems reasonable.

One can do this for a lot of major police departments. I tend to be skeptical of the tin foil hat people who say police departments are under reporting homicides. Maybe they are, but it accounts for a couple percent at most (pick your favorite place - Los Angeles, Baltimore City, etc and try it). The largest difference is accounted for by the smaller police departments who don't report at all.

"I tend to be skeptical of the tin foil hat people who say police departments are under reporting homicides"

For reference, see 'leftist conservative' below.

Just to be clear, there IS a big reporting bias in "crime" - caused by both cultural factors and police department pressure, among other issues. Rape for example, is widely considered underreported, and has been for a long time.

However, I am saying *homicide* is unlikely to be significantly underreported because its much harder to ignore bodies and we have an independent source (the CDC) of data.

"Rape for example, is widely considered underreported, and has been for a long time."

My wife, last night while watching TV, told me everybody knows it's overreported, because so many "rapes" are more cases of regretting it in the morning than non-consensual sex or violent coercion. I don't normally view her as having unusual views on things like this.

I know there are people who consider it underreported, but I thought they were extremists. Is this a mainstream view?

The view you and your wife have about rape is an extremist view with little connection to reality.

yes, this is a mainstream view among doctors (the AMA), bar association, law enforcement, and criminologists. The best available evidence is that people do not report it for the exact reason you bring up - people don't believe the accusation.

@GiT: the only extremist view is yours - the one expressed with no doubt and no evidentiary support.

I was a prosecutor for several years for a large US city and in the US Army. While all crime is despicable, rapes were the most difficult, emotionally and evidentiary, to deal with. Rape is BOTH the most overreported and most underreported of violent crimes. It is underreported because many victims fear reprisal, fear the ordeal of trial, or feel guilt or shame.

It is overreported because women know it is a powerful accusation against a man that can be supported by little more than her word and proof of sexual contact. A consensual encounter can easily be embellished. The allegation creates a shower of gushing sympathy toward the accuser. For the truthful accuser it is necessary comfort, and for the untruthful accuser, the attention is welcome. But they all initially get our sympathy.

There is a lot of Saturday morning regret, and a lot of women who initially claim rape recant. One cannot tell whether it was a true victim who backed down or a liar who feared getting caught. The latter is easier to guess: they slowly begin to unwind their story as their 'memory' is challenged by interrogation and evidence. The problem is that we do not often collect statistics on allegations that were withdrawn much less ascribe an informed opinion about the reasons for withdrawal.

Your opinion is just the knee jerk reaction of feminists who scoff at the idea that any woman would lie. The thousands of annual documented cases of actually catching women lying is the tip of the iceberg.

I think the FBI does a yearly phone survey where they try to get a handle on how often people are victimized in various ways, and I think they ask whether those crimes are reported to the police. That seems like a good way to get some kind of data to answer this question.

Anon 12:05 pm:

Just as a philosophical nitpick, surely the relevant question isn't which side of this question has the extremists on it, but rather which side is right.

NPR ran an interesting story in late 2005 about how rapes during the aftermath of Hurricane Katrina in New Orleans were grossly undercounted by the NOPD, which was largely refusing to take reports from victims.

If they refused to take reports, then they didn't investigate and we don't know how many of the allegations were false.

More Stories Emerge of Rapes in Post-Katrina Chaos
by John Burnett

Morning Edition, December 21, 2005 · Law-enforcement authorities dismissed early reports of widespread rapes in New Orleans during the lawless days following Hurricane Katrina. But a growing body of evidence suggests there were more storm-related sexual assaults than previously known.

Female victims, now displaced from New Orleans, are slowly coming forward with a different story than the official one. Two national crime-victims' groups have reported a spike in the number of reported rapes that happened to storm evacuees. The numbers are not dramatic, but they are significant when seen in light of the official number of post-Katrina rapes and attempted rapes: four...

>tin foil hat people

Suggesting that people might do something very easy, which is in their own self-interest, makes you a "tin foil hat person," eh? All righty then.

Also, enough with the straw man about "ignoring bodies." There is zero need to ignore them. We're talking about explaining them away.

No, some tin foil hat people actually suggest that the NYPD and Chicago PD do more than merely reclassify homicides, they "hide" bodies in the Hudson or lake Michigan, because the departments are infested with organized crime. In that case they are literally suggesting bodies are hidden not merely the crime reports.

I have no doubt that police departments would like to hide a lot of reports on homicides, and they have a strong incentive to do so, but its very hard to do it because homicides are highly visible and also tracked by news outlets and victim's families and hospitals as well. In the news article, the Detroit News found about ~4 out of 300+ (again, not counting justifiable homicides which are separately reported) - all of which would show up in CDC stats.

It depends. The incentives have changed.

Back in the 80s, police would overreport or report the most heinous suspected crime as support for higher departmental expenditures.

In subsequent decades as crime rates dropped, the incentive changed to taking credit for effective policing.

Im not saying that all crime reporting is suspect of unethical reporting by police for their own self-interest. Police identify violations of the criminal code based on their best interpretation of the events at the time of arrest. There can be and often is disposition of the case as something else without changing the initial, inaccurate judgment of the officer.

It is typically in the gray area surrounding good judgment that unethical reporting takes place. Occasionally, the misreporting is so severe that it statistically leaves the bounds of common error. But it is usually a whistleblower who reveals the problem.

There was a horrifying story in the LA Times a few years ago about how in the 1980s and 1990s there were at least five separate serial killers operating in South Central LA, who had murdered at least 100 crack whores amongst them ... and nobody at the time had noticed the poor women were victims of serial killers. One of them was the Grim Sleeper who was named and arrested after he restarted his career in the 2000s. The Grim Sleeper case caused the cops to look back at their records and discover that at least four other serial killers were at work in that era, and they had since been arrested in other states and had confessed to these L.A. killings without the LAPD ever noticing! Presumably, the LAPD is probably better at bureaucratic paperwork than a lot of police departments.

However, I don't believe that many of the confessed killings were not listed as homicides by the LAPD. Instead, they were just assumed to be part of the general crack war violence of the era.

Odd to me that AT chose to focus on "stand your ground" laws in his aside, as opposed to the quite obviously more problematic issue of police officers who legally kill people.

He mentioned 387 law enforcement justified homicides. But that figure should be suspect, especially relative to the smaller number of private justified homicides, since the justice system is biased in favor of justifying police violence but biased against justifying private violence.

It's made worse by the fact that the authors he linked to are confused about the treatment they're writing about. Much of the substance of those laws was already contained in the case law of the states that passed them (and some states that didn't). For example, there typically weren't changes in the duty to retreat with these laws. One thing that was definitely part of the treatment in all cases was the (often sensational) publicity the laws received.

Sounds as though the CDC definition ("injuries inflicted by another person with intent to injure or kill, by any means") should exclude felony-murder, vehicular manslaughter/involuntary manslaughter, as well as reckless/negligent homicides. Unless medical bureaucrats are using a sloppy definition of "intentional" and they mean to have a broader, fuzzier conception of homicide.

And of course these are all a rounding error compared to the Roe vs. Wade re-definition of justifiable homicide.

What he said!

Homicide to the coroner also means situations where someone dies due to not doing something. Someone else was involved in the death. It isn't simply a cop shooting someone, and it has no guilt attached. So the cops shooting someone, or a homeowner shooting a trespasser would not be a complete number.

Yes. A person hit by a car would initially be classified as a homicide by the coroner although one would naturally lean toward suspicion of murder with a firearm compared to death by car.

Coroners report 'cause of death.' Their autopsy report is going to go into a lot more detail, providing evidence that supports or refutes intent or specific theories in consultation with detectives and the DA.

Let's just say that their report to the state or federal government is not their highest priority. Their workload, especially on a murder case, can be gruelling. It's not all CSI neat and quick.

I once rescued from the middle of a busy intersection in Chicago a pedestrian who had been hit by a car going 30 mph . If she had died, it might have made sense for the coroner to list it as homicide to allow a police investigation. On the other hand, as I explained to the cops, the accident was overwhelmingly the fault of the young woman pedestrian who was sprinting across the intersection against a red light to catch a bus.

Exactly, especially if it was accompanied by fleeing the scene, another crime. We often think that a fleeing person was at fault in the accident, but in a surprising number of cases, that isn't true or at least their culpability is rebuttable by conditions.

A common occurrence is a jogger at night with dark, non-reflective clothing. A driver hits the person and, not knowing whether they are at fault don't stick around to find out.

Drunk drivers also flee scenes where they were not at fault. In one interesting case I had, a drunk driver fled the scene and drove to a bar where he "consoled himself" with alcohol. We did not get him on drunk driving, but he plead guilty to fleeing the scene, property damage, and other infractions.

are we pretending that crime stats arenot massively massaged and even faked by municipalities? Cuz I don't know if I can do that, even to be part of the crowd (which is very important to me!)

when are "educated" americans going to learn that americans are controlled/managed in large part by a skein of fake/bogus/made-up/massaged to a fare thee well statistics, data and studies? And that this skein of bogus stats is disseminated by the major societal institutions?

“injuries inflicted by another person with intent to injure or kill, by any means.”

That's a terrible definition for use by the CDC. There is no way to determine intent by the person signing off on a death certificate. They can suspect intent, but do they go back and correct the records based on the outcome of the police investigation.

Where does negligent homicide/manslaughter fit in these numbers?

Also, the CDC definition would include executions, i.e., excusable homicides.

All we can really tell is that they use the generic "homicide", the killing of a human by a human, while defining it more specifically, such as excluding the killing of a human by that same human, suicide, legally justifiable/excusable killings of humans, etc.

The reports are updated as time, information and motivation permits. In other words, they are not perfectly accurate.

"This section also includes information about justifiable homicide—certain willful killings that must be reported as justifiable or excusable. In the UCR Program, justifiable homicide is defined as and limited to:

1. The killing of a felon by a peace officer in the line of duty.
2. The killing of a felon, during the commission of a felony, by a private citizen. Because these killings are determined through law enforcement investigation to be justifiable, they are tabulated separately from murder and nonnegligent manslaughter. Justifiable homicide information can be found in Expanded Homicide Data Table 14, “Justifiable Homicide, by Weapon, Law Enforcement, 2006–2010” and Expanded Homicide Data Table 15, “Justifiable Homicide, by Weapon, Private Citizen, 2006–2010.” "

I'd venture a guess that most self defense shootings are not counted as justifiable homicides. For example, Zimmerman wasn't convicted of murder, but Martin was probably not committing a felony. Even though the jury decided that was a self defense shooting, it does not meet the criteria of a justifiable homicide. It's very possible that some of those unjustifiable homicides were actually in self defense.

"I’d venture a guess that most self defense shootings are not counted as justifiable homicides."

Going by definition #1, "The killing of a felon by a peace officer in the line of duty," then might it be the case that any number of self-defense shootings by policemen or other peace officers aren't justifiable homicides either given a reasonable inference of a deadly threat can be made before a felony is committed. I don't know the process that takes place after a police officer shoots someone, but you can't try someone who's dead so no determination of felonious intent can be made. The only determination that would seem possible is whether the officer was within his/her rights to use deadly force.

This is all pretty interesting because I'd always assumed that "justifiable homicides" included every killing where a determination had been made that use of deadly force was lawful rather than felonious status of the decedent or the act he or she was committing.

Trayvon Martin was on top of Zimmerman and was pounding his head into the concrete. There is no jurisdiction in the world where that would not be a felony or the equivalent. You just can't go around pounding people's skulls in because you think they are following you. The jury rightly saw it as an act of self defense.

It was a justifiable homicide by any sane reading of the events.

I wonder if this includes traffic fatalities. Things like hit and run drunk driving accidents etc. Accidentally killing someone with your car is just as much a homicide as killing them with a gun.

Free information found with google is worth every penny you paid for it??

And thing free from government is evil?

What data do you get when you pay the private sector as much as they demand/need to provide you with verifiably accurate homicide data?

Alex has been citing data from government agencies that have had their budgets attacked for doing wasteful studies and research and for collecting data with a leftist political agenda. Specifically, funding any research into gun deaths is prohibited by Federal law with Federal revenue because all such research is intended purely to violate the second amendment.

Just much federal funding is there on the damaged caused by legalized abortion?

How about slavery? Where is funding for research that will determine whether poor blacks would be better off as slaves?

Once something becomes recognized as a right, costs and benefits dwindle from the debate and research.

To take a couple of well-known deaths in my neck of the woods as illustration:

Nicole Brown Simpson, 1994: Homicide!
Michael Jackson, 2009: Homicide?

Just about every bureaucratic process that isn't wildly corrupt or negligent is going to count the former as a homicide, but it's likely that different processes will differ on whether or not to count the latter.

It would seem like a good topic for an academic to investigate is how much bias there is over time and over geography caused by these various methodological issues involved in reporting homicide. My offhand guess would be: not all that much, but I could well be wrong.

Good examples.

Aside from ethical integrity, investigative resources and expertise might explain anomalies. Coroners in rural counties, often conjoined with the Sheriff job, often lack the rigorous training and education of a big city medical examiner.

Given that the majority of homicides happen in urban areas, I have doubts about whether this would move the national statistics even though rural areas make up the majority of investigative offices.

In some places, COD is determined by a jury or inquest.

CDC gets funding for producing research on gun violence with the objective of adding evidence for the administration's gun-control initiatives. Individuals within the CDC have career incentives that are, I imagine, similar to NRA researchers.

I suspect the source of the distinction is that the crime reports are meant to comprehend murder and aggressive manslaughter. There are lower grades of manslaughter which are excluded from considerations but which are still species of homicide. In New York, to take one example, this would include 2d degree manslaughter (homicide due to reckless conduct, which is to say that a risk is known and disregarded), criminally negligent homicide (which is to say that the defendant failed to perceive a risk that a reasonable person would), two grades of vehicular manslaughter, and assisted suicide.


i would not necessarily call those inaccuracies .The incidence of “unfounded rape claims” was something like 5 times higher than the national average. Murder is a little tougher to hide. “Unfounded cases of murder” would be an interesting file to read. But, the incentive to fudge the numbers is overwhelming so they will find a way - See more at:

Every time I see a post here with no context, and if after finishing reading it I still have no clue what point the author is attempting to make, i know it's an AT post. Different sources offer different homicide data, so what? Same for almost any stat in this world.

Comments for this post are closed