< Calculating Body Counts


Friday, March 16, 2012

BROOKE GLADSTONE:  This is On the Media. I’m Brooke Gladstone.

BOB GARFIELD:  And I’m Bob Garfield. Kony is unquestionably a monster, but a true understanding of what kind of monster comes down to a methodical accounting. And in war crimes tribunals and truth and reconciliation commissions, international criminal courts, to say nothing of the court of public opinion, that means body counts.


The person most often charged with producing these body counts is statistician Patrick Ball, the go-to tabulator of crimes against humanity. Very simply, Ball’s approach is borrowed from ecology and the problems inherent in counting animals. He looks at death lists compiled by aid groups, governments, anyone really, and he charts the overlaps in the lists. Those overlaps reveal how representative a sample is. And knowing the accuracy of a sample enables him to calculate the total number of people killed, with real precision. Chart that against all other data available and he gets a real picture of who was killed and when. Patrick, welcome to the show.

PATRICK BALL:  Glad to be here.

BOB GARFIELD:  Before there was a you, what was the state of the art?

PATRICK BALL:  Well, people have been trying to figure out how many people died for a long time, People have done so by looking at transportation mechanisms in the Third Reich, how many people could have been taken to the death camps, for example. People have done demographic approaches, comparing different censuses and trying to figure out how many people are essentially missing. People have also done, for a long time, enumerations where you try to make a list of every person who is known to have died. Each of those approaches has strengths and weaknesses. None of them gets you to an estimate of true conflict-related mortality by violence.

BOB GARFIELD:  You actually left out one category, numbers being polled directly out of the air, such as by-


-Vietnam era U.S. generals who gave the press what they wanted, hard numbers, albeit not connected to any kind of reality.

PATRICK BALL:  I’m sure that some of those numbers were invented out of thin air. Many of them may also have been people’s seat-of-the-pants guesstimates, based on observed counts. And I think that that’s actually something we see a lot of today. People will say, well we know about 30 or 40 deaths in a particular event or in a series of events in a place but we think we’ve only seen half of them, so we’ll say there are 60 or 80. And I think that that mix of guessing and direct reporting is very common. What we want to do instead is mix direct reporting with a mathematically reliable estimate.

BOB GARFIELD:  Tell me, ultimately, why it matters whether 40,000 people were killed or 80,000 people were killed, except for the – victims themselves? How do wrong estimates do a disservice to the world at large?

PATRICK BALL:  Well, I’m not sure that it matters enormously whether the number is 60,000 or 80,000. But it matters a very great deal how those deaths are distributed. Let me give you some examples. In Peru, working with the Truth and Reconciliation Commission, my colleagues and I observed that there were approximately the same number of deaths reported that were attributed to the Peruvian Army and the Shining Path, the Sendero Luminoso, the Maoist guerrillas.


So if we had just concluded, on the basis of what was reported, that the two parties committed equal numbers of deaths, we would have misunderstood the deaths that were not reported. Once we corrected in Peru for the deaths that had not been reported, we realized that Sendero Luminoso was responsible for half again more deaths, and this dramatically changed the understanding of the conflict in Peru.


So it’s really important to get the numbers right, not for the total – that’s valuable and important, but more for the ways that we divide the total up over time, over space, by perpetrator, by the ethnicity of the victim, and so forth.

BOB GARFIELD:  The death toll numbers, of course, are not just a reflection of how many people died. They’re also frequently political weapons. Can you give me examples of how the numbers attached to certain conflicts have been abused by one interested party or another?

PATRICK BALL:  The one that comes to mind most clearly is the number of people who died in Bosnia between 1992 and 1995. During the conflict, it was often suggested, as far back as ’93, that 200,000 people had died. But I remember thinking after the massacre at Srebrenica that the [LAUGHS] number was still 200,000. And many thousands of people were killed at Srebenica, so how did the original number of 200,000 have any meaning?


Well, as it turned out, it didn’t really have any meaning. It was a number that was invented at a press conference and then cited and cited and cited, and through a process that people in my business call “citation laundering,” it became received wisdom. But when two different projects did two different, I think, very rigorous approaches to calculating the total, it became clear that the total number of deaths was about 100 to 105,000. That’s a big difference. And when those numbers were presented, people rejected them on these political grounds, that it couldn’t be so few.


Well, why not? What does the number 200,000 mean to people? And it’s interesting. We see this fairly frequently, the use of a number to signify a lot. We really don’t have a clear way in our heads to make seat-of-the-pants claims about numbers that are that big.

BOB GARFIELD:  Sometimes it’s not just the bad guys who are bandying about bad numbers. Sometimes it’s the ostensible good guys, like human rights groups trying to call attention to some horrible cruelty or another. And then journalists take that number and run with it.

PATRICK BALL:  Well, the first question we have to ask to a journalist who has received a number is, do you care if it’s right. I mean, does it matter that your story is accurate? I have had many of these debates with policymakers, with advocates, with journalists, and sometimes they say no, they don’t really care. Okay, well if you’re willing to fake the number, what are you not willing to fake? Are you willing to fake video, are you willing to fake a quote, are you willing to mislead your reader?

BOB GARFIELD:  They say they don’t care because the particulars are not relevant; it’s just a rough measure of the horror?

PATRICK BALL:  Journalists may defend a sloppy statistic because it’s close enough. They may defend it because that’s what their source said, because it seems plausible to them. When we hear a statistic, we assume that it has been done in a rigorous way. We think it has some kind of mathematical meaning. And if it doesn’t have that mathematical meaning, then I think we are misleading the people that we’re presenting it to. If a journalist is given a particular estimate of a total, it seems to me that they should contact an, an epidemiologist or a demographer or a survey statistician or a mathematical statistician, or even a social scientist to ask, is that plausible. And often enough, it may not.

BOB GARFIELD:  You are awash all the time in these statistics documenting deaths. Have you become inured to the human loss beneath the statistics? Have you lost or gained humanity in the process of making these numbers clearer for the rest of the world?

PATRICK BALL:  I don’t know if I’ve lost or gained humanity looking at these numbers for 20 years, but I make a big effort every time I look at a data set to remember that it’s not a data point; it’s someone who had hopes and dreams, was trying to accomplish something in their lives, had people they loved and people who loved them and their life was cut short by violence.


In Colombia several years ago I was presenting to a group that included family members of the Disappeared People, and

a woman came up to me after the conference and said, you know, I know about disappearances in my community that no one will talk about and that no one will ever document, and she said, so their names will be forgotten soon. At least they should be counted. That’s really the minimum we owe them, it seems to me, to remember that they existed.

BOB GARFIELD:  Patrick, thank you.

PATRICK BALL:  Thank you very much. It’s a pleasure.

BOB GARFIELD:  Patrick Ball is Chief Scientist and Vice President for Human Rights at Benetech in Palo Alto, California.