Florida State Rep. Matt Caldwell, R-North Fort Myers, appeared on CNN to explain why he and his colleagues voted not to consider a bill to ban assault-style weapons after the horrific high school shooting in Parkland, Fla.
During the interview, Caldwell drew upon a favored talking point.
"A law enforcement officer is more likely to commit a crime than a CWP (concealed weapons permit) holder," Caldwell said Feb. 21.
The comment caught our ear and we were curious about the data behind it.
We reached out to Caldwell and got no response. His source is most likely number-crunching by John Lott, head of the Crime Prevention Research Center, a group that has written about what it sees as the media’s anti-gun agenda, the need for parishioners to go to church armed, and similar themes.
In 2015, Lott compared the rate of crimes committed by police nationwide to permit revocations in Texas, Florida and Michigan. The numbers on police crime came from a national search of news reports by a team at Bowling Green State University.
Lott took the national data and compared it to the fraction of permit holders in Texas who had their permits revoked.
Lott worked the data several ways and found a substantial gap.
"The rate for police was between 7 to 10 times higher than for permit holders," Lott said.
But Lott’s work contains one clear counting flaw.
His data on crimes by police came from research by Philip Stinson, a criminal justice professor at Bowling Green State University. Stinson told PolitiFact that Lott failed to note that Stimson was reporting the number of cases, not the number of individual officers involved in crimes.
"Some officers are arrested more than once, and some arrested officers have more than one arrest case because of multiple victims," Stinson said. "So, the whole analysis is wrong because the rates are mangled by using a different unit of analysis."
Lott said he doubts that would make a sufficient difference.
"The number of officers committing crimes could be higher or lower, but you’d have to have it be 7 to 10 times lower in order to reverse the ranking." he said. "That doesn’t seem very likely."
There’s another counting issue.
Two leading researchers questioned Lott’s use of permit revocations to measure crimes by permit holders.
"Those data undercount the true number of infractions," said criminologist Richard Rosenfeld at the University of Missouri-St. Louis. "Think of the process: an infraction occurs, then it must be reported to the police or another legal authority, the police must find it as a crime or infraction that would warrant revocation, and then the revocation must occur."
Rosenfeld said instances fall out of the dataset at each step along the way.
Lott agreed that national victimization surveys show that relatively small percentages of crimes get reported, even fewer result in an arrest, and a smaller fraction yet lead to conviction.
"Revocations won’t capture the full number," he said. "I don’t know how to compare that pattern to police crimes because there are no matching survey data for police.
Lott said he believes that police would be less likely than the average permit holder to be caught in a crime, but that is more theory than proven trend.
Law professor and statistical analyst John Donohue at Stanford University gave another reason to doubt the link between revocations and crime. In his research, he found a "number of cases where the permit holder was killed while committing a crime and of course those permits are not revoked."
Donohue added that much research shows that when the number of police rises, crime goes down.
Caldwell said that a law enforcement officer is more likely to commit a crime than a concealed weapon permit holder. That is based on shaky numbers.
The article behind Caldwell’s statement treated the number of criminal cases involving officers as if it were equal to the number of officers themselves. According to the researcher that looked at those crimes by police, that "mangled" the crime rates of officers. A single officer could be named in many cases.
On the flip side of the ledger, other researchers said that using permit revocations undercounts the number of crimes by permit holders. The analyst behind the Caldwell’s statement agreed that it is an imperfect measure.
In short, the statement rests on dicey numbers. Those numbers point to the need for more research, but they don’t prove anything.
We rate this claim False.
Update: After this item published, we heard from John Lott and added his comments.