Earlier this week, Justin Elliott had a great piece at TPM Muckraker exploring how the notion that Umar Farouk Abdulmutallab, the Christmas Day bomber, had left Nigeria on a one-way ticket for Detroit—an idea that is not true—has gone largely unchecked in the media. As Elliot documents, reports that Abdulmutallab had purchased a one-way ticket, sourced to unnamed federal officials, first appeared on Christmas Day. Though correct reports that the ticket was actually a round-trip fare appeared in the American press no later than December 28, the one-way claim quickly became part of the narrative of the bombing attempt, and one of the key “red flags” that security officials purportedly missed in the run-up to the event. No less an authority than former attorney general Michael Mukasey referenced it in a January 6 op-ed for The Wall Street Journal; on Monday, it was repeated as fact in a McClatchy editorial.

Elliott calls this episode a “remarkable example of how bad information can travel far and wide,” which it is. But it’s also an example of how, in an integrated and free-flowing media environment, bad info can infiltrate the reporting even of news organizations that have previously gotten the story right.

Elliott flags one of these outlets: the Associated Press. The AP was one of the first places to report correctly that the ticket was round-trip, with that brief December 28 report, and got this point correct again in a longer story published on New Year’s Eve and a timeline of events published on January 7. But then, on January 8, the AP moved an article that included this statement:

Abdulmutallab apparently bought his ticket in cash, was flying the same day, had no check-in luggage and purchased a one way fare.

(A slightly different version of the story, with a different byline and dateline but the same language, appears here.)

So how does a news outlet incorporate a factual error after first getting it right? There may be a clue in the bylines and datelines. Several of the AP stories referring to a round-trip ticket came out of Nigeria, but Lagos-based reporter Jon Gambrell was not listed among the eight contributors to the January 8 piece.

That helps explain why the error wasn’t caught. But where did it come from? An AP review, initiated after CJR made an inquiry, found, according to spokesman Jack Stokes, that it was “based on reporting with the security analyst cited by name in the story—John Harrison.” Harrison may well have picked up the bad information from other, incorrect media reports. Apparently, so had the the AP staffers who didn’t flag his misperception before it crossed the wires. “The story should have referenced our own previous reporting on the subject,” said Stokes.

A variation on the same tale can be found at the Los Angeles Times. In its news reporting, the LAT never reported incorrectly that Abdulmutallab’s ticket was for one-way travel. But the paper did, on January 6, publish a letter by a reader who made that incorrect statement. The same day, veteran correspondent Andrew Malcolm published a blog post (since corrected) also repeating the one-way claim. Asked what his source for the claim was, Malcolm replied “early reporting”—meaning, probably, his general sense drawn from the early reports. (He did not recall specifically whether he’d heard the one-way claim in his own reporting.)

So what does all this mean? With respect to Abdulmutallab and airplane security, probably not that much: for those who see the Christmas Day episode as a grievous security lapse, the one-way ticket was only one small plank in the argument. (Indeed, the president, who’s well-informed on this point, thinks some changes are needed.)

But there are lessons here for error correction in the modern media world. As CJR’s Craig Silverman recently noted, with the old fact-checking systems more or less decimated, there’s an urgent need for news organizations to devise new internal strategies—and train their staff in them. That’s especially so when, as in a case like this, it’s easy to tap into Google and find “confirmation” for a claim that may not be true. (To be clear, there’s no evidence here that the AP and LAT have, compared to the rest of the industry, failed to do this. On this issue, these outlets performed better than many others.)

Silverman has also talked about the need to think of corrections as “content that can stand on its own,” so that they “truly become integrated within the river of news we hear so much about.”

Greg Marx is an associate editor at CJR. Follow him on Twitter @gregamarx.