Can we trust crime numbers?
U.S. Attorney General Eric Holder warned in his weekly video address last month of an “urgent and growing public health crisis” in America: The abuse of heroin and prescription opiates.
The overdose death of actor Philip Seymour Hoffman in New York on Feb. 2 had thrust the issue into the news, and Holder’s comments were intended to give the Department of Justice a voice in an important public conversation.
But the first statistic Holder cited as evidence of this pressing national crisis came from a stale federal study that found a 45 percent increase in heroin overdoses from 2006 to 2010.
The citation was emblematic: In the midst of a revolution where criminal justice data move at the speed of light, the pace of federal statistical studies can seem glacial, critics say.
“Using 2010 data to tell us about a heroin problem in 2014 is ludicrous,” says Rick Rosenfeld, a criminologist at the University of Missouri-St. Louis.
“I don’t think we know if we’re in the midst of a heroin epidemic. I do know there are localities where the numbers are up. But to use numbers from four years ago as evidence of an urgent national problem today is pointless and silly.
“It just shows you how primitive the crime information infrastructure remains in this country.”
OTHERS RELEASE REPORTS
The criticism often is focused on the federal Bureau of Justice Statistics (BJS), created by Congress in 1979 to collect, analyze and disseminate crime-related data.
Separately, the FBI compiles the annual Uniform Crime Reports — a seeming redundancy that is another point of contention for critics of federal statistics.
For 25 years, BJS functioned as a wellspring of national crime data. But over the past decade, its eminence has suffered as the media, universities and advocacy groups have begun releasing their own crime data sets and reports, often on a more timely basis.
As a result, journalists and researchers seeking current data and trends are turning more frequently to such organizations as the Transactional Records Access Clearinghouse, the Death Penalty Information Center and the Vera Institute of Justice, among many others.
For example, Pew has become a primary source on recidivism — a core question as states consider effective criminal justice reforms.
BJS touts its role as a source of statistical evidence for new “smart-on-crime” policies. Yet the relevance of its dated evidence is in question: BJS has not produced a new report on recidivism since 1994.
That finally is about to change. Within weeks, BJS is expected to release a fresh national recidivism report that a team led by the bureau’s Howard Snyder has been working on for years.
Insiders say the report has been delayed by attempts to standardize widely variable state records on recidivism. (One state might compile data on those who return to prison within three years, and another within five years.)
Snyder tells The Crime Report the new recidivism data is expected to be updated every year or so.
‘MOST CURRENT DATA AVAILABLE’
It is inevitable that national crime data is a snapshot of years past, of course. Most BJS reports come from data that is a year or more old.
“I often joke to reporters that when we look at the national data it’s like driving while looking in your rearview mirror,” says Melissa Sickmund, director of the National Center for Juvenile Justice in Pittsburgh. “You’re always looking at past years’ data. It is frustrating, but here’s the phrase you use: ‘The most current data available.’
“That’s just the way it is.”
To be fair, no one blames the overtaxed statisticians who work at BJS.
James Lynch, BJS director from 2010 through 2012, says the bureau has been hollowed out by funding cuts as a result of the 2013 federal budget sequestration, a hiring freeze and animosity toward the Department of Justice on Capitol Hill.
“BJS has been under-resourced for many years,” Lynch, now a criminal justice professor and department chair at the University of Maryland, tells The Crime Report. “If you want timely statistics, then bang on the door of your damned congressman.”
BJS has been subject to a Justice Department hiring freeze since 2011, and its 2014 budget of $45 million is unchanged from 2009, according to a bureau spokeswoman.
That budget is miniscule by federal standards. Its Department of Labor equivalent, the Bureau of Labor Statistics, has a 2014 budget of $592 million.
In one bright spot, BJS was recently authorized to fill longstanding vacancies by adding five statisticians to the 35 now on staff, Lynch says. (The spokeswoman says BJS employs 46 people total.)
“You can’t do science without scientists,” Lynch says. “If your staffing is down 40 percent, you are going to have to make choices about what gets done first.”
QUALITY TAKES TIME
Crime data flows to the federal government in a river that begins as a trickle of information from municipalities that may or may not follow the same standards — for example, the definition of felonious assault or the financial threshold at which a misdemeanor theft becomes a felony.
Sickmund and Lynch say the process of “cleaning” this divergent data to ensure standardization is a complicated, costly and time-consuming process that often sets the quality of BJS reports apart from that of non-governmental organizations.
“It takes time to get quality data,” says Lynch.
But what is the value of timeliness?
“When you don’t have any data, some data — reasonable data — is better than none at all,” says Rosenfeld. “When the agency responsible for key data is not disseminating on timely basis, others will fill the gap.”
Twenty years ago, the pioneering use by the New York Police Department (NYPD) of applied research to respond to crime problems created a “revolution” that continues today, says Rosenfeld.
The NYPD’s CompStat program and many like it across the country analyze weekly statistics, allowing commanders to react quickly by pivoting law enforcement strategies to address emerging trends.
Similarly, timely national statistics could provide comparative data to help local agencies understand the scope of a burgeoning problem — a surge in heroin use or carjackings, for example.
Lynch and Rosenfeld say the Justice Department should consider the data-release model used by the Bureau of Labor Statistics, which produces monthly reports on the job market, unemployment, earnings and consumer prices that are covered closely by the media.
The release of BJS reports is infrequent and irregular, and Lynch says the bureau might benefit from “the discipline” of meeting periodic fixed data deadlines like its labor counterpart. Of course, this would require funding increases that may not be forthcoming.
FBI MISSES REVOLUTION
Sickmund says criminal justice professionals are focused on improving the efficiency of data collection and dissemination. Her juvenile justice agency works diligently with state and local data-providers to enhance the quality of raw information.
But Washington funding drives the engine of national data collection. And if broad systematic improvements that speed processes are to happen, Lynch says, “the will must come from the federal government.”
He adds, “In the Justice Department pecking order, the Bureau of Justice Statistics and the Office of Justice Programs (its parent agency) don’t stack very high.”
And that is why no one offers much hope that the FBI — with its $8.1 billion budget and many friends on Capitol Hill — will be willing to give up its Uniform Crime Reporting Program portfolio, even though many view the UCR as a symbol of the nation’s crime-data inefficiencies.
Rosenfeld has long been a critic of the FBI’s pace of release of Uniform Crime Report data, typically about 14 months after the close of the pertinent calendar year. He says that makes the data largely irrelevant to law enforcement agencies.
A few years ago, he published an article in Criminology and Public Policy suggesting that the UCR be transferred to the BJS as “an important beginning step.”
He wrote:
“The United States does not monitor crime rates on a comprehensive and timely basis…The absence of up-to-date information on crime rates at the national and local level misinforms policy responses, impedes criminal justice planning and efficient resource allocation, and contributes to public ignorance of the crime problems in this country.”
Why do two separate Justice Department branches deal with statistics? In a word, politics.
The UCR was created in 1930, and its chief cheerleader was FBI Director J. Edgar Hoover, who used the crime survey (initially released quarterly) as proof that the country needed more FBI agents.
But 35 years after the creation of BJS as the country’s criminal justice data clearinghouse, many wonder why the FBI is still responsible for the UCR, which uses data submitted voluntarily by 16,000 law enforcement agencies.
Rosenfeld says the FBI’s unrushed pace in releasing the UCR each year serves as an example of the broader shortcomings of America’s crime information infrastructure.
“It seems as though the FBI is simply unaware of the crime data revolution,” says Rosenfeld. “The big UCR innovation in recent years (by the FBI) was to get it out in spreadsheet form.”
—-
David J. Krajicek is a contributing editor of The Crime Report and a veteran criminal justice journalist. A true crime author, he writes “The Justice Story” for the New York Daily News and is a 2014 fellow with the Fund for Investigative Journalism. He welcomes comments from readers.