behind the news

Sorry, Wrong Number

It’s not OK for journalists to be bad at math
November 20, 2009

Every year, Scott Maier, an associate professor of journalism at the University of Oregon, asks his students to raise their hands if they went into journalism because they love writing. Unsurprisingly, most of them put their hands in the air.

“Then I ask how many of them got into journalism because they love math and numbers, and the hands that stay up are pretty few,” he said. “In many cases, they got into journalism to stay away from math.”
Journalists love to joke about how we suck at math. “As mathematicians, journalists make fine geishas,” declared a correction in the West Australian last year. If you look at the sheer volume of numerical errors committed by the press every day, you’d probably be inclined to agree.

Almost every story contains a number, be it a statistic, an address, or someone’s age. Journalists deal with numbers every single day, and yet so many of us willingly profess ignorance or fear when faced with simple arithmetic. This fear combines with a lack of training to rank numerical errors among the most common mistakes made by journalists. I have an ever-expanding archive of these errors.

But it’s time to recognize that handling and interpreting math and numbers are some of the cornerstones of journalism. “As newspapers and news magazines chart a new course for themselves where they have to explain rather than just report, what’s happening is that our reliance on math and numbers becomes even more important, rather than less so,” Maier said.

Numerical errors usually occur for one of these five reasons:

• A journalist mishears a correct number given to them by a source and fails to double-check it.

Sign up for CJR's daily email

• A source unwittingly provides a mistaken piece of information and the journalist fails to verify it.

• A source deliberately fudges the numbers and the journalist fails to verify them.

• A journalist or editor miscalculates a figure.

• A journalist re-reports a mistake made by another media outlet.

These are very straightforward scenarios, all of which can be solved with two basic actions. First, journalists need to acquire the basic math skills needed to properly handle numbers and figures. Second, they need to develop the habit of double-checking every number and figure. (I’d call this the mantra of “Math twice, publish once,” but I don’t want to sacrifice language at the altar of arithmetic.)

Things become complicated—not to mention embarrassing and embarrassingly hilarious—after inaccurate numbers make their way out into the world. The more often a statistic is reported and re-reported, the more authority it assumes. At a certain point, an incorrect number gains a wake of supporting citations that can challenge or overcome the actual facts.

Another byproduct of numerical errors is amusing corrections. I hand out the Numerical Error of the Year every December, and am never short on entries. Last year’s winner was from The New York Times:

An article on Wednesday about the delivery of Barry Bonds’s 756th home run ball to the National Baseball Hall of Fame and Museum misstated the number of votes cast in an online contest held by Marc Ecko, the fashion designer who purchased the ball and asked people to vote on what to do with it. About 10 million votes — not “10,000 million” — were recorded.

Zeroes, and specifically how many of them to use, are a persistent problem. The Advertiser of Australia this week ran a very confusing correction in a similar vein:

LAST week’s NIE resource on Science questions answered by CSIRO incorrectly stated that there are 12 zeros in one billion billion. There are 24 zeros if a million times a million is doubled. Many now regard a billion as a thousand times a million. This would be 18 zeros when doubled.

(After seeing this correction on my Web site, a reader e-mailed to say the paper actually means “squared” when it says “doubled.”) That’s a mathematical error, but this week also saw The New York Times provide an example of a simple numerical error:

An article on Nov. 1 about libraries with rare-book collections open to the public misstated the period of time covered by Oscar Wilde’s college notebook, at the William Andrews Clark Memorial Library in Los Angeles. It was written during 1876 and 1878, not 1876 and 1978. And because of an editing error, the article rendered incorrectly part of the Latin title of Galileo’s “Starry Messenger,” at the Linda Hall Library of Science, Engineering and Technology in Kansas City, Mo. It is “Sidereus Nuncius,” not “Sidereus Nucius.”

Yes, amusing stuff. But the core issue here is learning to prevent these mistakes. We have to lay waste to the idea that it’s okay for journalists to suck at math. One reason is that this attitude makes journalists feel as though they can’t challenge the statistics contained in a press release. We’re not a fancy retweeting service for companies and governments—our role is to verify information. The other reason to shed the idea that journalism and math don’t mix is that it may, in fact, be completely false.

In 2003, Maier published an article about journalists and math in the Newspaper Research Journal. As part of his research, he gave a math test to reporters at the Raleigh News & Observer. When I interviewed him for my book, Maier told me that, despite the “profound feelings of mathematical inadequacy” professed by the editors and writers, most of them did fairly well. As Maier noted, “strong performers in math outnumbered the weak performers.” Yet nearly all of them were afraid of taking the test, and most expected to do horribly. It’s an admittedly small sample size, but it sends the clear message that journalists need to stop beating themselves up about math.

Fortunately, journalists can easily acquire the math skills necessary to properly do their job. There are some excellent, free online guides for journalists. This online presentation from CubReporters.orgprovides a solid overview of percentages, crowd counts, and means and medians, among other information. The great writing and reporting instructor Jack Hart also has several pages of excellent numeracy advice available online here. It’s a fantastic resource. Finally, Robert Niles has a page dedicated to math for journalists.

There are also a few books written specifically for journalists. The gold standard has for a long time been News and Numbers: A Guide to Reporting Statistical Claims and Controversies in Health and Other Fields. You should also look at Math Tools for Journalists and Numbers in the Newsroom: Using Math and Statistics in News.

Ah, but what of that remark about journalists, mathematicians, and geishas? Here’s the full correction, for your enjoyment (and calculation):

E=mc3+1: As mathematicians, journalists make fine geishas. One of the paper’s most perspicacious readers has again successfully challenged our careless checking of figures in reports received from overseas and interstate. In one report we had an Olympic swimming pool holding a meagre 1000 megalitres – a waist-high depth that would becalm Eamon Sullivan (‘Angel’, 4, drowns as plastic dam wall fails, page 17, November 25). And in another report we had 40,000 US “gleaners” filling 80,000 4-6kg sacks with 250kg of vegetables – a minuscule 6g per person (Hard times bite in America, World page 28, November 26). We still don’t know what we meant.

Correction of the Week

“ON 18 September 2009, we published an article in which Warren Furman, also known as the Gladiator “Ace”, was reported as denying “internet rumours” that he had raped Jordan. In doing so, the article implied that these “rumours” were sufficiently serious to require a response from Mr Furman.

“In fact the “internet rumours” consisted of very few ambiguous posts on an internet chat forum.

“They have since been deleted. We wish to make it clear that neither the posts nor any other matter, provided the basis for the false suggestions that Mr Furman may have raped Jordan. We apologise unreservedly to Mr Furman for the distress and offence caused to him by the article appearing to suggest otherwise.” — Daily Star (U.K.)

Craig Silverman is currently BuzzFeed's media editor, and formerly a fellow at the Tow Center for Digital Journalism.