Part 3—Narrative conquers all: Last Wednesday, the New York Times did a horrible terrible job with a full-length report on the gender wage/income gap.
The analysts were gnashing their teeth by the time they finished their reading. “Can’t anyone at the Times handle statistics?” one youngster even asked.
We felt the analysts’ pain. One day earlier, the Times had presented a full-length, front-page report about police shootings in Philadelphia.
The topic is very important and also very current. There too, we thought the Times struggled with some basic, important statistics about an important topic.
In our judgment, Matt Apuzzo didn’t do a terrible horrible job with his front-page report. We would say that his reporting was weak in some basic respects.
As he started, Apuzzo sketched a very familiar portrait of big-city police. He discussed a report by the Justice Department. Hard-copy headline included:
APUZZO (3/24/15): Report Details Deadly Forced Used by PoliceThe framework there is very familiar and very current:
Roughly once a week, 390 times over the past eight years, Philadelphia police officers opened fire at a suspect. The shootings involved 454 officers, most of them on patrol. Almost always, the suspects were black. Often, the officers were, too.
Fifty-nine suspects were unarmed. Officers frequently said they thought the men—and they were almost always men—were reaching for a weapon, when they were actually doing something like holding a cellphone.
In Philadelphia, police frequently “opened fire” on suspects. “Almost always, the suspects were black.”
Since the shooting of Michael Brown, this has been an extremely familiar framework in our public discourse. As Apuzzo opened his report, there it was again.
To state the obvious, Apuzzo was discussing a very important topic. Beyond that, he was discussing the findings of a Justice Department investigation—an investigation requested by Philadelphia’s police commissioner two years ago.
Apuzzo’s front-page report appeared in the Times last Tuesday. In the eight days which have passed, you haven’t heard squat about it.
Almost surely, that's because the Justice Department report on which it was based involved a bunch of actual facts. By way of contrast, our nation’s discourse tends to turn on bogus statistics, anecdotal incidents and highly familiar narratives.
Eight days later, let’s discuss Apuzzo’s report. If you read all the way to the end, we think it tends to contradict some widely-held impressions.
In the past few years, certain impressions about police shootings have taken strong hold. People have received the impression that police shootings are on the rise. Many questions have been raised about the frequency with which young black men are the target of these shootings.
We were intrigued by Apuzzo’s report because he was drawing on actual facts from an actual investigation. This may explain why you’ve heard nothing about his report in the past eight days.
Apuzzo started by painting a very familiar portrait. Much later in his report, he seemed to discard that portrait.
All too often, this is the way it tends to go when Times reporters work with statistics. Before we look at Apuzzo’s reversal, let’s consider several facts from his intriguing report.
In just his fourth paragraph, Apuzzo presents a striking statistic. He has already noted that the Justice Department report included “the kind of data that has been nearly absent from the debate over police tactics that began last summer with a deadly shooting in Ferguson, Mo.”
Then, he offers a striking comparison between the police departments in two of our biggest cities. New York City comes off rather well:
APUZZO: Only a handful of major departments regularly publish statistics on police shootings, and those that do are not always consistent. That makes comparing the records of police departments difficult. But even with such spotty figures, Philadelphia stands out when compared with the public data in other cities like New York, Los Angeles and Chicago. In many years, Philadelphia saw more police shootings than New York, a city with five times the number of residents and officers.New York is five times Philadelphia’s size. But in some years, New York has seen fewer police shootings!
Especially given the publicity surrounding such shootings in New York, that struck us as a startling statistic. Needless to say, it disappeared into the ether surrounding Apuzzo’s report.
What explains the high rate of police shootings in Philadelphia? As Apuzzo continued, he frustrated the analysts with several statistical claims.
He said Chief Ramsey requested the DOJ study “after a philly.com article in 2013 revealed that while crime was down, the number of officers firing at suspects was on the rise.” We’ll admit that we wondered how large that rise had been.
Later, Apuzzo said that Philadelphia “is the nation’s fifth-largest city but maintains the only police department that has more than 1,000 officers and does not offer a field training program for new officers.” We’ll admit we wondered if Philadelphia is the only such department out of three, or possibly out of 300.
(Near the end of his report, Apuzzo made a throw-away reference to “the violence plaguing the city.” Why did we feel certain that this plague has been on the decline?)
Those were minor points. Very late in his full-length report, we encountered a major claim.
Very late in his report, Apuzzo seems to contradict the familiar portrait found in his opening paragraphs. As he does, he offers a statistical assessment which struck us as somewhat odd.
You had to read all the way to paragraph 17 (out of 21). If you did, you encountered this striking assessment:
APUZZO: Over all, despite some high-profile cases of the police shooting African-American suspects, the report did not find a pattern of racial targeting. While blacks accounted for 80 percent of police shooting victims, whites who were shot by the police were more likely than blacks to be unarmed, the Justice Department concluded. When federal investigators looked at the race of the officers and the suspects, they found no statistically significant difference in the outcomes.“The report did not find a pattern of racial targeting?” Given the general shape of the discourse, we thought that assessment was striking.
We also thought it was somewhat odd.
The report didn’t find a pattern of racial targeting? That almost seems to contradict the familiar portrait with which Apuzzo opened his front-page report! On a somewhat brighter note, it seems to contradict a general picture of police conduct which has dominated public discourse in the past few years.
All around the national discourse, the impression has been given that police are targeting blacks, especially young black men. At least in Philly, the Justice Department seems to have reached a different conclusion.
We don’t know if that’s the case. This is why we found that assessment somewhat odd:
To us, it seemed odd to be told that there was no “pattern of racial targeting” even though “eighty percent of police shooting victims” were black. To us, that seemed like a high percentage.
This being the Times, Apuzzo didn’t say what percentage of Philadelphia residents are black. Incomparably, we decided to look it up.
According to the leading authority, Philadelphia was 44 percent black as of 2012. This basic statistic never appeared in Apuzzo’s full-length report. We’d say it adds a twist to the Justice Department assessment, which may be perfectly valid.
Philadelphia was 44 percent black—but 80 percent of police shooting victims were black! Despite those figures, the DOJ “did not find a pattern of racial targeting.”
According to Apuzzo, this benign assessment was based on the percentage of police shooting victims found to be unarmed. The benign assessment may be correct, but we’d make two points about it:
Again, that benign assessment flies in the face of a dominant narrative about police shootings. Indeed, Apuzzo seemed to advance that familiar narrative in his own opening paragraphs, the ones which appeared above the fold on the front page of the Times.
According to Apuzzo, the Justice Department found that Philadelphia police had not been behaving in that stereotypical way. But so what? At the Times, this finding was only available for people who read all the way to paragraph 17 on page A15, inside the hard-copy paper.
Let’s state a second point. In recent discussions of Ferguson, statistical disparities like these were taken as prima facie evidence of police misconduct. At pseudo-conservative news sites, people were told that such statistics can’t be taken at face value as evidence of police wrongdoing.
We pseudo-liberals received no such warnings from our own news orgs.
Ih Philadelphia, the Justice Department apparently found that a disproportionate pattern of shootings didn’t indicate targeting. But New York Times readers didn’t see a discussion of that, in part because Apuzzo omitted a key statistic—the percentage of Philadelphians who are black.
To what extent have police in Ferguson been targeting blacks in various kinds of police actions? We don’t know, but the Philadelphia example reminds us that surface disparities in these statistics may not always mean what they seem.
We don’t mean to suggest that Ferguson police haven’t been targeting blacks. It’s possible that Ferguson police have been targeting blacks to a greater extent than statistics seem to suggest.
We mean something different. We mean to say that our big news orgs should be examining statistics from places like Ferguson to gain a clearer idea of the way such cities work.
That will never happen, of course, given the nature of our journalism and our public discourse. Our discourse runs on narrative and on bogus statistics—and on exciting anecdotal events where the facts have often been massaged to generate “perfect examples.”
If Apuzzo’s report is correct, New York City police seem to fire their guns much less often that Philadelphia police. If his report is correct, Philadelphia police tend to shoot up the joint, but they apparently don’t target blacks.
Each finding flies in the face of widely-bruited perceptions. But you’ll never hear these findings discussed in our public square.
In the American public square, you’ll hear a lot of bogus statistics. Charles Blow will provide the latest anecdote—the latest perfect example.
You’ll also hear familiar narratives. Apuzzo seemed to open with one, then later seemed to reject it.
Philly police don’t target blacks? Apuzzo buried a striking assessment.
That said, the assessment flew in the face of a dominant narrative. And to the extent that we have a discourse, it very strongly tends to be narrative—narrative all the way down.
This afternoon: A much-bruited statistic about police shootings
Tomorrow: If it weren’t for misleading test scores, we’d have no test scores at all!
Friday: Dueling statistics on campus rape, and the narratives we liberals love