Back to my web page at

This is chapter five of Firing Back. Some graphs have been dropped from this chapter in the interests of speed. Yet another reason to buy the book!

What Crime Statistics Really Mean

Before we start throwing numbers around, it's important to know that the statistics about violent crime and guns come from a lot of different sources. In order to refute some of the nonsense that gets thrown around by the gun control proponents, you need to understand that many of the statistics that are available are much less reliable and meaningful than they appear.

Uniform Crime Reports program

One source is the FBI's Uniform Crime Reports (UCR) program. This program was started in 1929 to gather data from police departments across the United States about the most serious common felonies. Until 1959, however, many police departments didn't report crimes to the FBI, or did so sporadically. Even today, only 95% of the U.S. population is covered by law enforcement agencies that report to the UCR.[1] The only UCR data that is really of any value nationally is from 1959 forward.

There have long been objections about the accuracy of the data that comes out of the UCR program. The strongest concern is that police departments, under pressure to make their own city look good, may not be completely accurate in reporting crimes, especially if they have any discretion in deciding whether a crime was committed or not.

Another concern, also quite legitimate, is that many crimes are not reported to the police. As an example, a burglary where very little of value was taken might not be reported. Rape is another crime that has traditionally been underreported, out of shame, or a perception that the police aren't going to pursue it with any vigor. A victim who was breaking the law when attacked is also less likely to report the crime. For example, a drug dealer who is robbed of his merchandise is unlikely to report the crime.

A third charge is that many police departments give considerably more attention to crimes committed against whites than they do to crimes committed against blacks, especially poor blacks. At the same time, many of the same critics insist that police departments arrest and charge blacks with serious felonies far more often than whites.

national crime victimization surveys

Because of the many criticisms of the UCR data, the Department of Justice's Bureau of Justice Statistics started a regular survey program called the National Crime Survey (NCS). Unlike the UCR program, the NCS doesn't rely on victims to report crimes to the police. Instead, the NCS interviews about 101,000 people every six months to find out what crimes have happened to them. The 101,000 people involved are distributed around the country in a way that makes them representative of the whole population of the U.S.

Because it doesn't rely on the law enforcement agencies, there's no opportunity for crime figures to be "improved" for public relations reasons. Because it asks victims to identify characteristics of the criminals who injured them, it doesn't suffer from the problem of racial bias by police or prosecutors. On the down side, because murder victims have a hard time filling out surveys, it doesn't tell us anything about murder.[2] Because the surveys are done every six months, participants often forget about crimes that they reported to their local police. Blacks and poorly educated people also seem to have a habit of underreporting minor crimes to the NCS.[3]

But because the NCS surveys only 101,000 people, instead of tracking crime data from every police report filed in the whole nation, the NCS can ask very detailed questions like, "Did you fight back? What sort of weapon, if any, did you use? Were you injured?"

other surveys

There have been lots of surveys done about crime. Some are done by the same organizations that ask Americans for whom they intend to vote. Some are done by sociologists, criminologists, or public health doctors. Unlike the NCS and the UCR data, which have historically been free of biased questions or dishonest techniques for gathering data, some of these surveys are questionable. In some cases, the survey questions have been designed to get the "right" answer. In other cases, the people to be surveyed are very atypical of Americans, and the results may tell us very little about the rest of us.

not every homicide is a murder

When someone quotes figures for homicides in the United States, and compares them to other countries, make a point of asking them if they mean "murders and non-negligent manslaughters" or if they mean "homicides." Homicide includes any intentional killing of one person by another. This includes justifiable homicides, excusable homicides, murder, and manslaughters.

The way that international organizations gather public health statistics don't usually make distinctions between these various types of homicides. When gun control advocates cite public health statistics to prove how horrible of a society the United States is, make sure that the audience is aware that many of the "homicides" in the United States are defensive homicides by civilians and police, and that public health homicide statistics aren't always directly comparable from nation to nation.

self-defense killings with guns are underreported by the FBI

"justifiable homicides" aren't what you think

Gun control proponents like to quote an FBI statistic about the number of justifiable homicides done with guns each year. They may tell you how many handgun justifiable homicides there were, or how many justifiable homicides were done with guns by women. Make sure that you find out from where they get their number of justifiable homicides. If they get the numbers from the FBI's Uniform Crime Reports, make sure that you explain that "justifiable homicide" means something different to the FBI than it does when the rest of use that same expression. Even by the FBI's own definitions, the FBI dramatically undercounts justifiable homicides.

"Justifiable homicide" is a term with different meanings in different jurisdictions. Throughout the United States, a justifiable homicide is an event where one person kills another person to prevent a felony, and the law considers it legally acceptable. Everywhere in the United States, if you use deadly force to protect yourself or someone else from death or great bodily harm, you are within the law. Increasingly, many states are updating their statutes to presume that if someone breaks into your home, that they intend you great bodily harm, and therefore you are justified in shooting someone who breaks into your home.[4] If the person against whom you use deadly force dies, this is "justifiable homicide."

There are some special exceptions. As a general rule, if you provoked a fight with someone, refused to back down from the fight, and then killed the person, you are going to have a hard time getting the courts to call it "justifiable homicide."[5] In a few states, you are required to back away from an attacker as long as you can do so - even if it means leaving your own home.

In some states, especially in the West, justifiable homicide can include killing to suppress a riot, prevent arson of a dwelling that is occupied (or might be occupied), or to prevent a fleeing felon from escaping.[6] The courts, however, tend to take a very dim view of private citizens "playing policeman." Even policemen are severely restricted now from shooting a fleeing felony suspect because of the U.S. Supreme Court decision Tennessee v. Garner (1985).

excusable homicide

There is another category, called "excusable homicide." California's excusable homicide law defines it as one of two cases. The first case involves homicides "committed by accident and misfortune, or in doing any other lawful act by lawful means, with usual and ordinary caution, and without any unlawful intent."[7] This is the case where someone does everything right to go shooting, and someone else wanders into the middle of the range and gets shot. In other words, it's an accident.

But the second category of excusable homicide is so similar to justifiable homicide that you may not immediately see the difference: "When committed by accident and misfortune, in the heat of passion, upon any sudden and sufficient provocation, or upon a sudden combat, when no undue advantage is taken, nor any dangerous weapon used, and when the killing is not done in a cruel or unusual way."[8] If someone ran up to you on the street, knocked you to the ground, and you pulled out a gun and shot them, this would be an excusable homicide. Why isn't it a justifiable homicide? Because you may not have been in danger of death or great bodily harm. On the other hand, excusable homicide laws recognize that under the circumstances of "sudden combat" you don't have time to make that subtle distinction.

civilian legal defensive homicides

Gun control advocates often point out that justifiable homicides with guns are only 1.4% of all murders, and therefore, guns are not an effective method of self-defense. It is certainly true that the FBI's Uniform Crime Reports only showed 308 civilian justifiable homicides with guns in 1992.[9] But there is a considerable body of evidence that shows these figures greatly understate the total number of cases in which a criminal is shot dead by a civilian. If you want to know how many civilians kill criminals each year, you need to look at not only justifiable homicides, but also the "sudden combat" excusable homicides. Let's call this combination of civilian justifiable homicides (CJHs) and the "sudden combat" excusable homicides "civilian legal defensive homicides" (CLDHs).

One reason for underreporting is that the FBI makes the distinction between "justifiable" and "excusable" homicides that we discuss a few paragraphs back. The distinction is very subtle - but the police record these distinctions, and the FBI only reports the justifiable homicides, not the excusable homicides.[10] Another problem is that police reporting of CLDHs is done less carefully than police reporting of murders.

There is no data available for the United States as a whole that tells us the yearly number of CLDHs. But there have been several studies of different cities and counties that tell us how many CLDHs there are, relative to the number of murders and manslaughters. The noted criminologist Gary Kleck has concluded, based on these studies, that the number of CLDHs with guns per year is typically 7.1% to 12.9% of the murder rate (at least five times the FBI's "justifiable homicide with a gun" figures).[11]

"Seven Deadly Days" & CLDHs

There is another problem with the FBI's figures for justifiable homicide that significantly understates CLDHs, and significantly overstates murders. If the police investigate a homicide and ask the district attorney to charge someone with murder or manslaughter, that is reported as a murder or manslaughter to the UCR program. But district attorneys often investigate a case, find evidence that the killing was indeed, justifiable or excusable homicide, and drop the charges. A person who is tried for murder will sometimes be found innocent because the killing was done in self-defense. This is very often the case in spousal abuse situations where a woman defends herself or her kids from a current or estranged husband with a gun.[12] If a murder turns into a CLDH after the initial report has been taken, there is a strong possibility that this change won't make it into the UCR data.

How do we find out how many such cases there are? We have an especially interesting source of information, because it was originally produced as a piece of antigun propaganda. In 1989, Time magazine ran an article called "Death by Gun." It included photographs and information about every person killed by a gun in one week in the United States.

The week was May 1-7, 1989. Was this a typical week for gun deaths in the U.S.? Reasonably so. There were 464 gun deaths reported in the article. Of these, 216 were reported as suicide, 14 deaths were CLDHs, 13 deaths were police justifiable homicides, and 22 deaths were accidents.[13] This leaves 199 murders.

To scale up the May 1-7 gun deaths to determine yearly rates for the United States, we can't just multiply by 52 weeks per year. Murder rates peak in the summer months; May, 1989, had 7.8% of 1989's murders.[14] To scale up May 1-7 gun deaths to an annual rate, we multiply by 100.0/7.8 (May's murder percentage), and then multiply by 30/7 (the fraction of May days that included May 1-7). This gives us the following results (remembering that the "per year" figures on the last line are extrapolations):

Seven Deadly Days     total gun     suicide CLDHs    police  accidents murders 

initial report        464           216     14       13      22        199

initial report        100.00%       46.55%  3.02%    2.80%   4.74%     42.89%  

per year              25,495        11,868  769      714     1,209     10,934

For the year 1989, the FBI reported 11,832 gun murders, 236 civilian justifiable homicides, and 360 police justifiable homicides.[15]

That the FBI's 1989 figures for gun murders, civilian justifiable homicides, and police justifiable homicides don't match our extrapolations isn't all that surprising. Only about half of all police justifiable homicides are actually reported to the FBI;[16] the data above fits well with this fact. The 769 CLDHs are far higher than the FBI's figures for civilian justifiable homicides, but we've already discussed why the FBI's numbers are too low.

We talked a few paragraphs back about Professor Kleck's estimate that CLDHs with guns should be 7.1% to 12.9% of the total murder rate. For 1989, that would be between 1,346 and 2,445 gun CLDHs - not 769.[17] Remember that the Time article, like the FBI's reporting, showed the number of civilian defensive uses initially reported. A year later, Time decided to follow up on the murder cases, and see how the courts handled them. Instead of 14 CLDHs, now there were 28 - 14 of the murders reported in "Death by Gun" were now ruled justifiable homicides.

Seven Deadly Days     total gun     suicide CLDHs    police  accidents murders 

one year later        464           216     28       13      22        185

one year later        100.00%       46.55%  6.03%    2.80%   4.74%     39.87%  

per year              25,495        11,868  1,538    714     1,209     10,165

This number of CLDHs (1538) is at the low end of the range that Professor Kleck's estimates would give. However, at least 43 murder cases had still not gone to trial,[18] and it was still possible that some of these would be found "justifiable." It does seem unlikely, however, that most of these cases still awaiting trial one year later would be found to be justifiable or excusable homicides.

The most likely explanation for the projected CLDH count being at the low end of Kleck's estimates is that the studies upon which Kleck based his estimates were all conducted in urban counties. It is logical to expect in urban counties, where crime rates are highest, that there would be fewer CLDHs, even relative to the overall population.


1. Federal Bureau of Investigation, Uniform Crime Reports 1992, 289.

2. Uniform Crime Reports 1992, 386.

3. Gary Kleck, Point Blank: Guns and Violence in America, (New York: Aldine de Gruyter, 1991), 175-176.

4. California Penal Code §198.5 is an example of such a statute.

5. California Penal Code §197 is one example.

6. California Penal Code §197 allows deadly force "to apprehend any person for any felony committed, or in lawfully suppressing any riot, or in lawfully keeping and preserving the peace." The courts (at least in California) have narrowed the section about apprehending fleeing felons to just the crimes that were felonies under English common law (e.g., murder, rape).

7. California Penal Code, §195.

8. California Penal Code §195.

9. Uniform Crime Reports 1992, 22.

10. Uniform Crime Reports 1992, 22.

11. Kleck, 111-114.

12. Kleck, 114.

13. "Seven Deadly Days", Time, July 17, 1989, 30-60.

14. Crime in the United States 1991, 14.

15. Crime in the United States 1992, 18, 22.

16. Kleck, 114.

17. Crime in the United States 1992, 18. There were 18,954 murders in 1989.

18. "Death by Gun: One Year Later", Time, May 14, 1990, 30-31.