386. Go figure
Two divisions of the Department of Justice have released crime figures this month. The press release from the Bureau of Justice Statistics, reporting the results of the National Crime Victimization Survey, says:
In 2008, an estimated 4.9 million violent crimes (rapes or sexual assaults, robberies, aggravated assaults and simple assaults) occurred...
That's 1,930 victimizations per 100,000, of course. Compare that to the figures released by the FBI 12 days later:
An estimated 1,382,012 violent crimes occurred nationwide in 2008... ["Violent crimes" are defined as "murder and nonnegligent manslaughter, forcible rape, robbery, and aggravated assault."]
There were an estimated 454.5 violent crimes per 100,000 inhabitants in 2008.
The chief explanation is that little word "simple." About 3,472,590 of the assaults estimated by the NCVS are classified as "simple." The survey defines simple assaults as those "requiring less than 2 days of hospitalization", a period of hospitalization that might strike some as non-trivial.
In 1980, according to the CDC, the average length of hospital stay for all conditions was 7.5 days, while in 2006 it had shrunk to 4.7 days (see table 102 on page 392) [researching blogs ain't for sissies], so I wonder if the two-day cap might be a bit outdated. Among medical insurers, one day is the new two days, but our crime statisticians might not have kept up.
Both the NCVS and FBI basically define aggravated assault as an attack with a weapon, although the wiggle room built into their definitions differs somewhat (NCVS here and FBI here.) To stir in a bit more confusion, the term is used to designate an act that in many criminal codes, and in the tort system everywhere, is called battery and not assault at all.
Another difference between the two agencies' estimates is that the FBI counts homicides, a class of crime that has proved resistant to the NCVS's information-gathering techniques.
If the "simple" assaults are subtracted, and homicides added, the NCVS figures would work out to 1,720,812 violent crimes, a difference of 338,800 extremely serious violent crimes. That's a lot of suffering to lose at the margin.
But the difference is still larger than that figure suggests, because as far as I can make out the FBI figures includes crimes against people younger than 12, who are excluded from the survey for obvious reasons.
And even so I think it's almost a given that the NCVS understates the prevalence of crime, because it depends on what respondents are prepared to say when called up. Would an abused 13-year-old speak freely to the voice on the other end of the line, when the abuser is waiting for his or her own turn on the phone? Would the battered wife describe the beating she's ashamed to reveal to her best friend, when her husband is sitting at the kitchen table a few feet away?
The NCVS reports that 41.4% of all rapes and sexual assaults are reported to police. A 2007 national report focused specially on drug-facilitated rape found that only 16% of rapes were reported to law enforcement, and estimated an annual total of over a million, or more than 4 times the NCVS number.
The amnesiac power of certain drugs was vividly illustrated to me by the orthopedist who set my son's broken arm. Both bones were completely broken - he looked like Elastic Man - but about five minutes after the horrible grinding setting the doc asked him, "Did that hurt?" and Alex said: "No." The doctor looked at us and said, "It's kind of scary." More than kind of, although I'm glad he was spared the traumatic memory.
When you start factoring in rapes facilitated with drugs like that, you have to wonder about the accuracy of any estimate that doesn't include pharmaceutical factory output.
Wednesday, September 16, 2009 at 09:05PM in Crime statistics
Reader Comments