Journalistic innumeracy, cont.

I've written before about journalists and statistics, but a couple of recent articles in America's paper of record, the New York Times, make it worth banging this particular drum a little bit more.

First of all, Rick Lyman, the Hollywood reporter, weighed in on Monday with an article about the relative box-office successes of Spider-Man and the new Star Wars film. Here's the relevant bits:

"Star Wars: Episode 2 — Attack of the Clones" sold an estimated $116.3 million worth of tickets in its first four days in theaters, an impressive opening for any film, but one, it must be said, that did not quite match the $115 million that "Spider-Man" made in just three days two weeks ago... Fox had tried to inoculate itself from comparisons to "Spider-Man" with a news release on Thursday pointing out that "Clones" would be on only 6,000 North American screens, far fewer than the 7,500 showing "Spider-Man." But even taking that into consideration, the $31,769-a-screen average of "Spider-Man" was far higher than the $27,254 for "Attack of the Clones."

It starts off with the genuine studio estimates: although the Star Wars numbers were subsequently revised downwards by a substantial $6 million, the blame for that must lie with Fox and not with Lyman.

But then comes this sentence saying that Star Wars made $27,254 a screen over 6,000 screens. If that was true, it would have grossed an astonishing – I daresay impossible – $163,524,000 in its opening weekend. They just gave us the actual gross, which was $116.3 million: can't they see that their numbers don't add up?

And the "per-screen" numbers are obviously meant to give us an opportunity to compare like with like, so they wouldn't give us a comparison between the three-day gross for Spider-Man, on the one hand, and the four-day gross for Star Wars, on the other. In fact, the average for Star Wars was based on its Friday-Sunday gross. But that's automatically going to put it at a disadvantage: since the real fanatics will have all seen the film on Thursday, the averages exclude the first (and, presumably, the most important) day of release.

But Lyman made a much more important mistake than that: he confused the per-theatre average with the per-screen average. Any Hollywood reporter – hell, any moviegoer – knows that blockbuster films are opened on many different screens in any given theatre. In fact, the per-screen average for Spider-Man, based on the initial studio estimates, was $15,312 – $5,104 per day – while the per-screen average for Star Wars, excluding the initial Thursday, was $14,358 – $4,786 per day. Including the initial Thursday it was $19,383 – $4,846 per day, a difference from Spider-Man of only 5%. When you take into account the fact that Thursday was a school day when people are less free to go to the cinema, the averages become essentially indistinguishable.

There are other factors one could take into account as well, such as the fact that the producers of Star Wars were pickier than most about where their film could be shown, which presumably meant that Star Wars was shown, on average, on screens with more seating than Spider-Man was. Those sorts of numbers are unavailable. But even so, to say that Spider-Man's averages were "far higher" than those for Star Wars is definitely off the mark.

Then, today, the New York Times publishes an article with the headline "Perelman Sells Stake in Bank to Citigroup for $5.8 Billion". The first sentence reads: "The financier Ronald O. Perelman is selling his largest and his most profitable holding, a stake in Golden State Bancorp, as part of a $5.8 billion acquisition of Golden State by Citigroup, the companies announced yesterday." The sentence is right; the headline is wrong. Citigroup is paying $5.8 billion for all of Golden State, not just for Ron Perelman's 30-ish per cent stake. If the headline had said "Perelman Sells Stake in Bank to Citigroup in $5.8 billion Deal", it would have been ambiguous but arguably correct. As it stands, however, it's indisputably wrong. The headline says that Perelman is trousering $5.8 billion, when in fact he's getting maybe $1.75 billion. It's a big difference: more than $4 billion.

I've dealt, once, with the New York Times editing process, and from my experience and that of friends of mine, I can tell you that it's very rigorous. Rigorous with grammar, that is. When it comes to numbers, New York Times editors are obviously asleep at the wheel. After all, both of these articles were internally inconsistent, and in neither of them was the mistake picked up.

Posted by Felix at 20:05 EST