When NYPD cops fired 50 times and killed unarmed Sean Bell a year ago, they called it “contagious fire.” When they shot and killed troubled teen Kheil Coppin last month, mistaking a hairbrush for a gun, it was spun as “suicide by cop.”
Now, thanks to a RAND Corporation study that it commissioned, the NYPD has a surprising explanation of the huge and disproportionate number of stop-and-frisks of black and Latino New Yorkers. What it boils down to is the RAND report’s conclusion, which the institute banners as this: “NYPD Pedestrian Stop Patterns Mostly Racially Neutral, with Some Trouble Spots.”
The RAND report was released just before Thanksgiving to mostly positive reviews in the press. However, the report is hardly an independent analysis. Not only was it commissioned by the NYPD, but its scope was specifically limited by the police, an interview with the study’s author reveals.
Prompted years ago by criticism in the wake of the shooting death of Amadou Diallo, the NYPD started requiring its officers to fill out reports on their stop-and-frisks. Just this year, the NYPD released a year’s worth of stats under pressure. After an outcry about what the stats revealed, it hired RAND to analyze the figures.
Make of the RAND conclusion what you will, but the raw data released in the report can leave many New Yorkers feeling raw: In 2006, city cops reported making 508,540 pedestrian stops, ostensibly to question people suspected of criminal behavior—yes, half a million. Fewer than 10 percent of the stops resulted in arrests or summonses, according to the NYPD’s own stats.
A stunning 89 percent of those stopped were minorities, including 267,000 black people. In other words, 55 percent of the stop-and-frisks were of black people—more than double their percentage of the city’s population. A total of 145,000 Latinos were stopped.
Not that this is any surprise to black or Latino New Yorkers, but 45 percent of those who were stopped were frisked, compared with only 29 percent of whites. Police used force about 50 percent more often on blacks than whites—despite the fact that, percentage-wise, cops found guns, drugs, or stolen property on whites about twice as often as they did on black suspects. That suggests less rigorous search standards used by cops against white people.
Whatever term you choose—racial bias, selective enforcement, racial profiling—the numbers seem to support the notion that the NYPD unfairly targets minorities, especially blacks.
No, says RAND statistician Greg Ridgeway, who conducted the study. Arguing in statistician lingo that there are unconsidered variables that skew the results and actually lead to paradoxical conclusions, Ridgeway’s study determines that the racial disparities of police stops are “much smaller than the raw statistics would suggest.”
In fact, he says, the argument could be made that black New Yorkers weren’t stopped enough.
As Ridgeway puts it, black New Yorkers were “understopped by about 20 percent relative to their representation of criminal- suspect descriptions.”
The RAND study, for which the department’s Police Foundation paid $120,000, concludes that no major changes should be made in the NYPD’s stop-and-frisk policies. There’s nothing paradoxical about that to the NYPD’s critics.
Christopher Dunn of the New York Civil Liberties Union calls the report “not only a whitewash, but a racist whitewash.” Dunn, the NYCLU’s associate legal director, adds: “To suggest that stopping blacks in large numbers is acceptable because more blacks commit violent crimes is outrageous, especially when one realizes that 90 percent of those stopped are law-abiding people who were neither arrested nor given a summons.”
THE CONTROVERSY OVER THE NYPD’S STOP-AND-FRISK statistics dates back to the 1999 Diallo case, in which the unarmed black African immigrant was gunned down in a hail of 41 shots while reaching for his wallet after police tried to stop and question him in the Bronx.
As part of the continuing fallout from that incident, the City Council required in 2001 that the NYPD provide quarterly reports detailing the racial breakdown of stop-and-frisk reports. The police department provided reports for 2002 and data for the first three quarters of 2003 but then stopped. In November 2006, the NYCLU formally requested that Police Commissioner Ray Kelly comply with the law. There was no response from the NYPD until it learned last January that The New York Times was preparing to do a story about the department’s refusal to comply, according to a lawsuit that the NYCLU recently filed.
This past February 2, on the eve of the story’s publication, the NYPD gave the City Council the stop-and-frisk reports for 2006. Those reports showed that the number of stops had quintupled over 2002, the last full year from which data were supplied, and that blacks were being stopped at a rate twice their percentage of the population.
To quell the subsequent shitstorm, Kelly announced three weeks later that RAND would analyze the NYPD’s data.
RAND released its report last month. But the report did not include all of the NYPD’s raw data, so there’s no way to check Ridgeway’s work or to see how else the numbers could be crunched.
And the NYPD has, of course, refused to release the data. The NYCLU, after attempting to obtain the stop-and-frisk database provided to Rand through a Freedom of Information request, filed a lawsuit in November to try to pry the information from the NYPD. Similar requests from the City Council have also been turned down by the NYPD. The NYCLU’s suit is pending.
Each study of the data would of course be different, but there were peculiarities about how RAND conducted its study. Unlike the reviews that RAND conducted of other police departments, most notably in Cincinnati and Oakland, its analysis of the NYPD sought no comment from community organizations, politicians, those who believe they were unjustly stopped, civil-rights groups, or the police union.
Had Ridgeway spoken with the Patrolmen’s Benevolent Association, for instance, the union officials would have told him that officers are required to reach quotas of stop-and-frisks, which could very well explain the huge number of stops. Each such stop gets entered on a form known as the “UF250.”
“It’s our view that the UF250s are an important tool, but once they’re subjected to a quota they not only become illegal, they become dangerous,” says PBA spokesman Al O’Leary.
In the past, Kelly has denied that the NYPD has quotas of any kind but says that it does have productivity goals. He has said that the difference between goals and quotas, which are illegal in police work, is that officers aren’t punished for not meeting the goals. O’Leary contends that there are quotas and that punitive action is common for not meeting them. “It happens all the time,” he says. “We end up defending our people on it all the time. It’s a real problem.”
Even Ridgeway’s study, which goes soft on the department, notes the pressure on officers to do more stops of people. The stop-and-frisk reports are tracked and evaluated at the NYPD’s weekly CompStat meetings, at which commanders are grilled about their precinct’s crime statistics. “NYPD’s CompStat focus,” Ridgeway’s report says, “gives officers a strong incentive to generate UF250s. An officer’s UF250 numbers suggest productivity. A precinct captain can use UF250 numbers to show that the precinct’s officers are active in the areas that are generating complaints and where crimes occur.”
Despite that acknowledgment, the RAND analyst says he spoke with only one officer about the issue. He says he was told that although supervisors push officers to file UF250s to ensure that cops aren’t loafing, there are no hard quotas.
In addition to analyzing the data, Ridgeway interviewed seven training officers and a handful of recruits to see if they understood the stop-and-frisk policy, and then was escorted by police for an eight-hour tour watching cops write up UF250s.
Ridgeway spent the rest of his time crunching the stats. Critics see the lack of comment and evidence besides the stats as a fatal weakness of the RAND report. “From the beginning,” says Dunn, “the police department excluded the community from any participation in the project. As a result, this report raises many more questions than it answers and will do nothing to help sooth police-community relations. I think the police department really missed a big opportunity here.”
And that apparently was the NYPD’s intention. Ridgeway praises the collaborative process in Cincinnati and Oakland, saying, “I think it was very effective.” But he adds: “That’s just not the approach the NYPD wanted to take with this report.”
The RAND analyst, who is based in California, seems taken aback by the stinging criticism of the report but equally surprised by the predominantly positive spin the media gave it when it was released in mid-November. “The NYPD should not take these numbers to indicate that there are no problems,” he says.
For instance, RAND flagged 15 officers who they found to be “seriously over-stopping minorities.” Those cops should be immediately reviewed and then monitored, he says. The racial disparities in the stops made by Staten Island cops “stood out in just about every outcome”: frisks, searches, arrests, and force used. The police department has to address that problem, Ridgeway says.
And he insists that even though his study found that the racial differences in those numbers weren’t as glaring as the raw data suggest, “there still are differences. But they’re on the order of five percentage points, as opposed to twice as much.”
Ultimately, Ridgeway declares that his report “does not absolve the NYPD of the need to monitor the issue, but it also implies that a large-scale restructuring of NYPD [stop-and-frisk] policies and procedures is unwarranted.”
THAT CONCLUSION OF RIDGEWAY’S COULD BE SEEN AS unwarranted if you look at the stats in the report. Not all of the startling stop-and-frisk numbers are explained away.
Even using “the most liberal assumptions” about the national average when it comes to the rate of the public’s contact with police officers, the study notes, New York should have had “roughly 250,000 to 330,000 stops rather than the 500,000 stops actually recorded.” The report states: “The number of stops appears to be quite large.”
Here are some other highlights of the report:
Given the generally positive news coverage that greeted the report, perhaps these negative findings got lost in the translation. The report relies on a methodology that leaves some professors who deal in this type of research scratching their heads.
RAND ruled out comparing population-to-stop ratios because the census “doesn’t indicate who is out there on the street,” Ridgeway tells the Voice. Population numbers, he says, also fail to take into account participation rates in crime.
Many of the results in the RAND study are difficult to compare with other reports, but the findings in general are in contrast to a December 1999 report by then–Attorney General Eliot Spitzer. That study, which analyzed 15 months’ worth of stop-and-frisk reports, about 175,000 in total, found that based on respective arrest percentages, there was a huge racial difference in stops.
Eight years later, the RAND report—which doesn’t include the supporting raw data—says that’s no longer the case, that the racial differences in the figures are practically negligible.
Another benchmark that RAND compared to the stop ratio was suspect- description figures. That test was the basis for Ridgeway’s statement that black pedestrians were actually being “understopped,” and it’s also the NYPD’s favorite fallback line. But the stop-to-suspect description analysis is based only on descriptions for violent-crime suspects. Suspect descriptions for other crimes were not available to him, Ridgeway says.
Having ruled out the census comparison, RAND determined that, based on the arrest and suspect-description benchmarks, black pedestrians were not overstopped.
The majority of the RAND analysis relies on an obtuse formula of “similarly situated” cases to assess if there were differences in the post-stop treatment—frisking, searches, summons, arrests, and use of force—for blacks, whites, and Latinos. Ridgeway describes his method in the report as trying to compare “pedestrians who were stopped in similar situations, in the same places, at the same times, and for the same reasons.”
Columbia Law School professor Jeffrey Fagan, who was a lead researcher on Spitzer’s stop-and-frisk report, says that in a city like New York, where many of the neighborhoods are racially segregated, figuring out numbers of “similarly situated” groups can be difficult. “You have to wonder what is being lost,” he says of the numbers crunched in the RAND report.
In one of RAND’s comparisons, for example, the pool of blacks similarly situated to white suspects was only 25,679, or 9.6 percent, out of the approximate 267,000 who were stopped by police last year. In another, RAND used only 8,260, or about 15 percent, of the approximately 55,000 stops involving whites they found similarly situated to black-pedestrian stops.
“If you lose too many cases, you have what is called a ‘selection bias’ in your sample,” Fagan says—and that can distort the results.
Officials at the Center for Constitutional Rights point out what they believe is a major shortcoming in the RAND analysis: “It fails to acknowledge and factor in that the majority of stops are based on subjective criteria and are not tied to alleged criminality.”
Noel Leader, the co-founder of 100 Blacks in Law Enforcement Who Care and a recently retired sergeant who spent 20 years on the NYPD, says cops call that “going fishing.”
“If you stop 50 people,” he says, “somebody’s going to have something—a bag of weed, a warrant, a gun, something you can arrest them for. But my training taught me we’re hunters, not fishers. Hunters go after specific targets: people you think are doing crimes. Fishermen throw a net out there and whatever he catches, he catches.”
“Going fishing” is something that wouldn’t reveal itself in the data, says Ridgeway.