Last year, I wrote a story for Smithsonian about white-nose syndrome, the fungal disease that’s killing cave-dwelling bats in the eastern United States. Researchers told me about watching sick, confused bats flutter out of caves in the middle of winter; about entering caves literally carpeted with bat carcasses; about picking bat bones, as slender as pine needles, out of their boot treads.
When I reported the story, scientists and wildlife managers estimated that a million bats had died since the epidemic began in early 2007. Baseline data were scarce, and the number was acknowledged to be little better than a guess. Now, after a long process of soliciting expert opinion and extrapolating from existing data, the U.S. Fish and Wildlife Service has announced a new estimate: between 5.7 and 6.7 million bats have died from white-nose syndrome — some 85 percent of all cave-dwelling bats in infected areas.
Agency director Dan Ashe called the roughly sixfold increase “startling new information.” But for most of us, it’s not. A million is a big number, and six million is a bigger number, but our minds aren’t very good at grasping either of them. Like the gazillion-dollar federal debt, the bat death toll is just big. And that makes it very easy to ignore.
In a well-known 2004 study, researchers described two Amazonian tribes whose members don’t have exact words for large numbers (their language instead refers to “one,” “two,” and “many.”) Tribal members also struggled to match groups of more than three or four objects, suggesting that specific language is key to understanding larger numbers. A later study, which conducted a similar test among speakers of informal sign language in Nicaragua, supported this conclusion.
No matter our language, psychologists have found that most of us can’t “subitize” — instantly assess the quantity of — more than four or five objects at a time. (Take a subitizing test here.) We have to count them, with words. At some level, we understand any number larger than a handful as an abstraction, even when it stands for real people or animals experiencing terrible things. It’s a statistic. And statistics, as the University of Oregon risk researcher Paul Slovic puts it, are “human beings with the tears dried off.” Or, in the case at hand, bats.
In a 2007 experiment, Slovic and two colleagues demonstrated the numbing effects of statistics: They gave each of their subjects a chance to contribute $5 to the organization Save the Children, then gave them three scenarios. In all three scenarios, subjects were told that their money would go toward relieving the African food crisis. In the first scenario, they were also told that their money would be earmarked for a seven-year-old Malian girl named Rokia. In the second, they were simply given statistics about food shortages in Africa. In the third, they were told that their money would go to Rokia, but they were also presented with food-shortage statistics. As expected, donations under the first scenario far exceeded those in the second scenario, but first-scenario donations were also significantly greater than those in the third — suggesting that the process of statistical numbing starts as soon as the number of victims exceeds one.
We know this: It’s the power of poster children. It’s why the World Wildlife Fund uses a panda bear as its mascot. And it’s why journalists try to connect with readers by telling a larger story through an individual. But for this approach to evoke sympathy, the individual has to be at least a little bit appealing. And bats are, well, an acquired taste. Do you instinctively want to hug this little guy?
I thought not. Bats are fascinating, in appearance and habits, and the more one knows about them the easier they are to appreciate. But they’ve got a cuteness deficit.
Where cuteness fails, analogies can help bring numbers down to earth. Six million is, of course, a number with weighty historical echoes. It’s also the number of people who would fill the Indy Motor Speedway 15 times over. And it’s three times the population of Houston. David Schwartz, the author of the children’s book How Much is a Million? likes to make sense of large numbers by translating them into time: “One million seconds comes out to be about 11½ days. A billion seconds is 32 years. And a trillion seconds is 32,000 years,” he says. “I like to say that I have a pretty good idea what I’ll be doing a million seconds from now, no idea what I’ll be doing a billion seconds from now, and an excellent idea of what I’ll be doing a trillion seconds from now.”
When it came to the bats, though, I found only one way to feel the numbers in my gut. I saw plenty of dead bats during my Smithsonian research, and they saddened me, but I was essentially a tourist, in the caves for only a few days. What I remember most aren’t the bats themselves, but my conversations with the researchers who’d spent days, weeks, and months walking over tens of thousands of bat carcasses. I heard the lead in their voices, and saw the looks on their faces. That was worth at least six million words.
Flash credit: iStockphoto.
Townsend’s big-eared-bat photo: Courtesy U.S. Forest Service.
Aw, he’s totally cute!
I talk to over 3,000 people a year about bats. Young people think bats are cute. Lots of mature people also think bats are cute. Videos of an orphaned leaf-nosed bat made Cute Overload.com and Huffington Post on the same day and crashed the originating organization’s website from the traffic. I think repeating ‘bats are ugly’ is no help whatsoever in this fight. While analogies are great to help kids wrap their heads around big numbers, letting people see what a content, relaxed bat’s face goes a long, long way toward engendering sympathy. A sleeping big ear bat with its ears folded down does not do a beautiful animal justice.
The new number is still basically just a guess. Censusing of bats is typically done with raw counts–a practice that is generally not methodologically sound, and precludes a valid statistical treatment (it is impossible to calculate valid confidence intervals for the estimate or to test a hypothesis that there is a decline because the assumption of 100% detectability is almost always violated). Further, although there have been instances of mass mortality, most of those “dead” bats are actually missing and presumed dead. This is a problem because populations are not monitored at a scale appropriate to the hypothesis that WNS is causing a general decline in population of hibernating bat species. There are hibernacula where censused populations have increased and there are almost certainly many hibernacula that are unknown–many of the apparently missing bats may actually have relocated rather than died. We don’t actually have a methodologically and statistically defensible estimate of how many bats are missing from known hibernacula or of how many of those are actually dead. They have replaced an old guess with a new guess–what we need to do is replace guessing with science.
Thanks, Lenny. What would be a better method of estimating mortality, given the lack of baseline data?
We reflexively want to monitor when bats are concentrated for winter hibernation because it seems efficient and cost effective and raw counts are used because of the sheer numbers of bats. However, in the last decade or so new statistical methods have been developed that would permit us to monitor bats in summer when they are dispersed (this is the appropriate scale for the question and it minimizes the impact of purely local mortality that can occur in hibernacula for reasons other then WNS). Presence/absence data collected acoustically (by recording bat calls) at 100 sites spread out in a large area with 3 nights recording per station (or route) would permit us to explicitly account for variation in detectability and allow rigorous statistical analysis that would permit an estimate of abundance with valid confidence limits and to test the hypothesis of population decline at at an appropriate scale (an entire state or a geographic region rather than a point).