Call it The Washington Post's 766-word non-correction correction.

Call it The Washington Post's 766-word non-correction correction.

It ran Saturday in the form of a Paul Farhi article about the dubious nature of trying to measure the size of Rush Limbaugh's radio audience. Farhi stressed that trying to determine the total number of weekly listeners represented an exercise “in guesswork, slippery methodology and suspect data.”

The article detailed how there aren't any hard ratings numbers within the radio industry regarding Limbaugh's audience size -- a topic of increased interest since the AM talker had emerged as the public face of the Republican Party under the Obama administration.

Farhi helped put the ratings question issue in proper context, but the unspoken point of the piece seemed to be to walk back the previous day's Post article by Howard Kurtz, which boldly announced in the very first sentence that Limbaugh's ratings had “nearly doubled” since the recent controversy with Obama began in January. It was a pro-Limbaugh proclamation that went off like a firecracker, especially online, as conservatives cheered the news and mocked Democrats for padding Limbaugh's audience.

But was it true? Reading Farhi's detailed ratings piece on Saturday, it was hard to see how Kurtz's claim of Limbaugh's audience doubling could withstand serious scrutiny. (A point I made the day the Post published Kurtz's piece.) In fact, on Saturday we found out the thin sourcing that Kurtz used for his claim was little more than a hunch, and the person who made the hunch didn't think the guesstimate of Limbaugh's size growth was scientific or that it represented a true ratings estimate. (More on that below.)

Plus, even Limbaugh himself didn't think his ratings had doubled in recent weeks. The same day Kurtz's article appeared, Byron York at The Washington Examiner asked the turbo talker about his ratings. This was Limbaugh's response, in full:

The latest numbers I have are for January, well before this kerfuffle began, and they are through the roof -- six shares in NY, for example. There are d aily ratings taken now in about the top 15 markets but I have not seen them yet. All I can tell you is that as of January, we booked 80 percent of all our 2008 revenue and we'll be over 2008 by the end of this month.

Limbaugh hadn't seen ratings more recent than January. Yet the Post on Friday claimed Limbaugh's ratings had nearly doubled since January. During his Monday broadcast, Limbaugh reiterated that he had no idea if his ratings had recently increased twofold.

Have Limbaugh's numbers spiked in recent weeks? I'd be shocked if they hadn't given the extraordinary publicity he's received. But doubled? There's simply no proof, regardless of what the Post claimed on Friday.

Kurtz's sloppy reporting highlighted the media's perpetual soft spot for Limbaugh's ratings. As Farhi noted, nobody has specific numbers about what the talker's audience is and “Limbaugh himself has muddied the water with the claim that he reaches 20 million people a week, although there's no independent support for that figure.”

Yet, for years, news consumers have been told 20 million people listen each week. It's a statistic that has become absolutely synonymous with Limbaugh.

But where did that ginormous number come from? From Limbaugh, of course. The first reported reference I could find came from the July 31, 1993, issue of the radio bible, Billboard magazine, which reported “Li mbaugh's show is now heard on 610 stations and reaches approximately 20 million listeners, according to [Kit] Carson,” Limbaugh's “chief of staff.”

According to Limbaugh's right-hand person, the talker had 20 million listeners. Was there any way to confirm that? Not really, but no matter: The media loved the nice round number, and soon it began to appear everywhere -- but often without the acknowledgement that the stat came from Limbaugh's camp. The following month, in August 1993, U.S. News & World Report announced: “Welcome, one and all, to Rush World, the one-man media theme park of the '90s. Over here, the Radio Show, reaching 20 million listeners a week on 616 stations.”

And the month after that, conservative columnist Cal Thomas wrote that Limbaugh “is heard on more than 600 mostly AM radio stations with an audience estimated at 20 million listeners per week, is a phenomenon unseen in modern times.”

A check of Nexis today finds more than 800 news references to that mythical Limbaugh number throughout the years. Despite the fact nobody knows if it's accurate, the figure has been codified: Limbaugh attracts 20 million listeners each week. Wow.

But what other type of media reporting do journalists gladly repeat ratings numbers based on nothing more than a feel-good estimate provided by broadcasters? (Why stop at 20? Why not claim Limbaugh has 50 million listeners each week?)

And how amazing is this: Limbaugh in 1993 claimed he had 20 million listeners, and in 2009 the press is still mouthing the same statistic. Meaning that, until recently, Limbaugh's audience hadn't budged -- not up, not down -- in 16 years.. Obviously that doesn't pass any kind of smell test.

Why is it so difficult to pinpoint the number? First, much of radio's ratings methodology remains stuck in the 1960s, and it takes months to generate nationwide audience figures -- unlike TV ratings, which can often be measured within 24 hours. And second, because Limbaugh appears on a patchwork of stations all over the country, it's tough to add up all the numbers for an accurate reading. As Farhi noted, Arbitron, the overseer of U.S. radio ratings, has never tried to measure Limbaugh's audience. And it has no plans to since, as its spokesman told the newspaper, “There is no economic motivation for any objective third party to do that kind of analysis.”

Obviously radio syndicators have ratings numbers off of which they sell advertising, but those figures are closely held -- unlike Arbitron data, which is more widely avail able. So, basically, it's up to the syndicator to dole out the ratings numbers to the press; like last year, when Limbaugh's syndicator, Premiere Radio Networks, claimed 20 million listeners tuned into Limbaugh's show each week.

Also keep in mind that eight-digit number is what's known in radio as the “cume” (short for cumulative). It in no way reflects the actual audience size like the way television shows are measured minute by minute or half-hour by half-hour. Instead, the cume number represents a very large -- and generous -- umbrella covering the number of people who, in theory, tune into a program at any time during the week, even if it's for just two minutes.

As a radio trade reporter confirmed to MSNBC last week, common industry shorthand to determine the actual size of a radio audience at any given moment is to cut the cume figure down by a factor of 10, which would mean Limbaugh's 20 million becomes 2 million. Or, if you take the more modest cume number of 14 million, which some inside the industry have used to judge the talker's audience, Limbaugh's rating becomes 1.4 million, which is roughly the same size audience that Rachel Maddow and Keith Olbermann get each night on cable TV. So why doesn't the press treat them as the ultimate kingmakers?

The point is, over the years the press seems to have gone out of their way to puff up Limbaugh's ratings and audience size, which brings us back to the Kurtz assertion that Rush's ratings have doubled. What was Kurtz's sole piece of proof for his sweeping declaration about Limbaugh's audience having doubled? Here it was:

“The people who love him are a very small segment of the public,” said Michael Harrison, publisher of Talkers magazine, whose research indicates that Limbaugh's weekly audience has spiked from 14.2 million to about 25 million since the controversy escalated.

A couple points about Harrison. His estimate, which Kurtz presented as fact -- as research -- was just a hunch; nothing more and nothing less. It was Harrison's best guess. As he told the Post in the subsequent Farhi article: "[I]t's what we're hearing, based on the e-mails, the calls, all the buzz this controversy is generating. We put a little bit of our interpretation on it, added it all up, and that puts you in the ballpark."

The fact that Kurtz used that tea leaf-reading methodology as the anchor for his claim -- in the first sentence of his article -- that Limbaugh's ratings had “nearly doubled” was rather remarkable, since there's absolutely nothing substantive to back up the claim.

In fact, over the weekend, Harrison expanded his thoughts in emails to AOL blogger Tommy Christopher regarding the latest estimates of Limbaugh audience size:

They are only our thumbnail estimates based upon our contacts in the field, tracking of Arbitron estimates and understanding of the business. We make no claims as to “scientific” accuracy. ... [T]hey are not “ratings” per se.

Also, note that Harrison's guesstimate for the Limbaugh audience spike applied only to last week, and even Harrison didn't think the alleged audience jump would last. Kurtz, though, made it seem as if Limbaugh had permanently doubled his audience since January.

The third point is that Harrison is an unabashed cheerleader for talk radio. That's his job -- to drum up business for the format. (Harrison last week: “All things taken into consideration, now is a great time to be a talk radio advertiser.”) And that's fine. Journalists, though, ought to take his ratings estimate with a grain of salt. But not Kurtz: He presented as “research” Harrison's guesstimate that Limbaugh's ratings had nearly doubled.

But the point here isn't Harrison. It's Kurtz, because he's the one who typed up the unequivocal lead to the Post article. And it was Kurtz, the high-profile media reporter, who people then cited as the author of the Rush-doubled-his-audience claim. And coming from the pages of the Post, that gave the bogus claim instant credibility.

• “Since the first of the year, Rush Limbaugh's audience has exploded, according to Howard Kurtz of The Washington Post, even as his daily assaults on Obama have intensified.” [D.C. Examiner]

• “Rush Limbaugh's Ratings Explode” [Free Republic, March 6.]

• “The more pressing matter is whether he can move beyond his base of listeners, which has reportedly doubled since the current controversy exploded.” (The New York Times' Opinionator blog, March 6.)

• “Limbaugh's numbers are doubled.” (Chris Matthews on The Chris Matthews Show, March 8.)

The truth is journalists only have a faint idea of how many people listen to Limbaugh's program each week. And until reporters can get some independently verifiable information, they shouldn't pretend hunches represent facts. And they shouldn't announce Limbaugh's audience has doubled unless they can prove it.