Are Murders Going Up or Down?
With all the publicity regarding nearly every killing of one person by another, it is reasonable to ask, is this crime getting worse in our country, or better? There are two national databases that collect relevant information on this. Why two? Do they agree with one another?
The Federal Bureau of Investigation has been collecting Uniform Crime Reports since 1930, so they have gotten very good at it. The UCR data comes from voluntary police reports, so it attempts to collect all and only crimes that are known to police. That data, for crimes that include murder, involve a lot of specifics: jurisdiction, circumstances (argument, robbery, gang-related), victim-offender relationship, and so on.
The FBI’s UCR data has two divisions for killings, one for all murders plus nonnegligent manslaughters and one for negligent manslaughters. Negligent manslaughter means only unintentional, accidental killing of one person by another. The first division includes everything else: justifiable homicide, murders, law enforcement-related killings, and killings in self-defense.
The second national database on death is managed by the Centers of Disease Control and Prevention (CDC). It is called the National Vital Statistics System (NVSS) and collects information by Fatal Injury Reports. Its data about deaths comes from death certificates filed by people in each state, primarily funeral directors, medical examiners, or coroners. The death certificates include information on the manner of death (homicide, suicide, unintentional) and also information about the cause of death (injury, poisoning, etc.).
CDC’s NVSS data covers virtually every death, because death certificates are filed for everyone who dies. The FBI’s UCR, coming from voluntary reports, has some omissions where police reports were not sent to the FBI, or where crimes were never reported to the police.
A bigger difference than the number of deaths in these two databases is the way they classify homicides. The UCR takes a predictably legalistic approach, classifying homicides as intentional, justifiable, or negligent. The NVSS puts all person-by-person death in one category, homicide, without judgment of criminal intent, because death certificate data cannot make those distinctions.
The U.S. Bureau of Justice Statistics just published a report on these two measures of homicide, describing in more detail than I have done here the methods and limitations of each system. Then the report includes a graph of the homicide rates by each method, over a 30 year period, 1981-2011. Here is that graph.
You don’t have to remember all of your school mathematics to understand this. Looking at each organization’s own definition of homicide, the graphs are striking similar. The NVSS rate is slightly higher, probably because that captures all the deaths while SHR misses some due to non-reporting by police.
The really striking thing is that in 1981 the homicide rate was 10 per 100,000 people in the US. It dipped to 8 in the mid 1980’ and climbed to a high of 10 again in 1991. Since then, for the last 20 years, the homicide rate has been falling or steady. Falling or steady to its 2011 level of 5, half that of 1991.
Murders are going down. Down, not up. Down, despite all the publicity being given to both mass and individual murders. Down.
If you want to see the beautifully short (4 page) report from which this material was taken, here it is. It contains pointers to many other reports and sources of data for you to explore if you are a data junkie, as I am.
The Nation’s Two Measures of Homicide NCJ247060 July 2014
Picture Credit: hoyou / Shutterstock