Conservative pundit Heather MacDonald is on a one-woman crusade to prove a war on cops that doesn't exist, while grossly exaggerating crime upticks in a handful of cities as a new crime wave.
She wrote: "The blog FiveThirtyEight looked at homicide data from most of the nation's 60 largest cities in September and found a 16 percent rise."
From that citation you'd think they had given MacDonald's assertion a ringing endorsement. She doesn't tell readers the article was headlined, "Scare Headlines Exaggerated The U.S. Crime Wave." Of course, the scare headlines in question have often come from MacDonald.
FiveThirtyEight found that a handful of the 60 largest cities saw homicides increase, most did not. MacDonald's claims were based on the handful of cases, as if she went out looking for cities where data supported her theory and ignored those that didn't.
FiveThirtyEight wrote "changes [in homicide rates] aren't statistically significant" for the vast majority of the 60 largest cities. They urged people to remember that homicides and general crime have dropped to record, or near record lows over the past few decades. When a number is quite low to begin with, even a small numerical increase looks large--if you only talk about percentages.
In her most recent fear piece in the Wall Street Journal MacDonald claims "An 11% one-year increase in any crime category is massive." No, it is neither massive nor insignificant by itself. It only takes on meaning when you know the base number. For instance, according to FiveThirtyEight, Tucson saw homicides drop 12%, but that was only 3 fewer murders. In Raleigh homicides dropped by 23%, but that too was just 3 fewer homicides. Anaheim saw a massive 44% increase in homicides, but that means 4 additional murders.
FiveThirtyEight wrote, "most cities have not shown a significant increase in homicides over the last year."
Earlier this year the Washington Post published an article entitled "In major cities, murder rates drop precipitously." They noted that since 1990 homicides in New York City dropped 85%, in Chicago the "overall crime rate has declined to its lowest rate since 1972." "The number of violent crimes committed per 100,000 people has been dropping even longer, from a high of 758 in 1991 to 367.9 in 2013. The rate hasn't topped 500 per 100,000 people since 2001."
After having published pieces pushing the claim that a new crime wave was coming, including pieces by MacDonald, the New York Times took the unusual step of publishing an editorial repudiating them. They acknowledge, "The rate of violent crime, including murder, has been going down for a quarter-century, and is at its lowest in decades. On average, it is half of what it was in 1990, and in some places even lower."
Since crime rates have fallen so substantially for decades even numerically small increases appear significant in percentage terms -- one reason MacDonald concentrates on percentages alone. The Times wrote, "Two lessons emerge from this data. One is that when crime rates are so low, even small changes can appear large. The other is that small sample sizes based on arbitrary time frames are nearly always nonrepresentative."
Fellow conservative John Lott, Jr., at the Crime Prevention Research Center, disputes MacDonald's claims. Comparing it to past claims, that also proved false, he says MacDonald's claim is "possibly more inflammatory" and "appears to be less accurate."
Patti Davis, Ronald Reagan's daughter, a liberal, echoes MacDonald. She claimed gun crime increased dramatically since her father was shot in 1981. In fact, the opposite is true. In 2013, Pew Research reported, "The gun homicide rate in 2010 was the lowest it had been since CDC began publishing data in 1981." Since 2010 homicides have continued to decline to even lower levels.
The FBI Uniform Crime Report for 2014 listed gun homicides in the U.S. at 8,874, with a total homicide count of 13,164. They dropped further in 2011, went up slightly in 2012, then dropped again in 2013. By 2014, the last year for which we have data, total homicides were down to 11,961--a drop of 1,203 or 9% in four years. Gun homicides followed a similar pattern and by 2014 had dropped to 8,124--a decline of 750 or 8.4%.
The Washington Post noted, "In 1993, there were seven homicides by firearm for every 100,000 Americans, according to a Pew Research Center analysis of data from the Centers for Disease Control and Prevention. By 2013, that figure had fallen by nearly half, to 3.6 -- a total of 11,208 firearm homicides." [Note: While CDC data and FBI data differ both show the same trend line over time.]
If the FBI stats are correct, or even close to the mark, with 8,124 gun homicides, and a population of 318.9 million, the gun homicide rate per 100,000 would have dropped to 2.5 in 2014. In the last quarter of a century gun homicides have dropped by two-thirds, from 7 to 2.5.
Conservatives like MacDonald, and liberals like Davis, both want crime stats to prove their bigger political agendas. MacDonald may want people to stop criticizing police, Davis may want gun control, both are using a bogus explosion of crime to push a political agenda.
As the editors at the New York Times warned: "Misunderstanding crime rates -- or worse, using them for political purposes -- makes it hard to have an informed debate about which policies will be most likely to keep violence down."