In the lead-up to the United States’ 2003 invasion of Iraq, Judith Miller’s byline regularly graced the front pages of the New York Times. An aggressive and bold reporter, she had cultivated a bevy of highly placed sources both in Washington and in the Middle East. Miller appreciated earlier than most the rise of global jihadism—a militant interpretation of Islam calling for confrontation with the perceived enemies of the religion—and its potential marriage with more destructive technologies. When a group of hijackers crashed planes into New York City’s World Trade Center in 2001, she was one of the few who had appreciated such a possibility. She was also well suited to report on the fallout.

Alongside President George W. Bush and members of his administration, Miller trained her sights on Iraqi strongman Saddam Hussein and the threat of weapons of mass destruction (W.M.D.) in Iraq. More than a decade earlier Hussein had chemical bombed his own people. She broke story after story, bolstering the case for invasion with scary details of Iraq’s actions and capabilities: renovation of storage facilities to house dirty radioactive bombs, or the possible illicit purchase of the smallpox virus for biological warfare, and the import of aluminum tubes for use on nuclear warheads.

The Bush administration was staffed with neocons—conservatives who aggressively pushed for the promotion of democracy and U.S. national interest in international affairs, including by force. Miller often relied on highly placed officials within the Bush administration, leading some to accuse her of acting as the administration’s neoconservative mouthpiece. “While Miller might not have intended to march in lockstep with these hawks, she was caught up in an almost irresistible cycle,” Franklin Foer wrote in a retrospective in New York magazine. “Because she kept printing the neocon party line, the neocons kept coming to her with huge stories and great quotes, constantly expanding her access.”

What contributed to her getting her central story so spectacularly wrong, though, was her reliance on anonymous sources. Not only was Miller parroting the administration’s party line, she based much of her reporting on Iraq on avowed foes of Saddam Hussein, who had skin in the game. In other words, they had a vested interest in inducing America to invade Iraq so they would be able to take power, or so they believed.

Embedding in a U.S. military team after the invasion, Miller watched firsthand as the search for W.M.D. came up empty. Yet, citing unnamed sources, she continued to file stories such as “Illicit Arms Kept Till Eve of War, An Iraqi Scientist Is Said to Assert” (April 21, 2003), based largely on third-hand information. In essence, Miller, and thus the Times, had been played.

The question, in this case, was one of accountability. Named sources are accountable to their words in a way that anonymous sources are not. Readers can understand a source’s agenda more fully when that source is properly identified. Naming sources enhances the pressure to speak truthfully, as in the event that a named source proves unreliable, the source is identified on record as having provided faulty information.

Nearly every newsroom prefers naming sources wherever possible. The New York Times expresses its “distaste for anonymous sourcing” in its standards and ethics guide; NPR has a strict code limiting what anonymous sources can say in stories, disallowing the use of sources offering “anonymous opinions on others.” Newsrooms do not ban such sources outright, though, because they can be essential for certain stories, particularly those involving sensitive, classified information. Sometimes freedom from the pressure of speaking as oneself allows an unnamed source more rope to speak honestly.

The question of anonymity raises accountability questions of another sort: who are journalists accountable to? The obvious, and primary, answer is to their readers. But they are accountable to their sources as well: to portray their words accurately and honestly. Sometimes, this accountability to sources means protecting their identity, when sensitive information is involved and reprisals are possible.

Still, too much leeway for anonymous sourcing allows for sources to spin stories in their preferred direction without the reader being any the wiser. That’s why New York Times deputy executive editor Matt Purdy calls them potential “journalistic I.E.D.s” (improvised explosive devices, or bombs), which can blow up unexpectedly and damage a story’s—and publication’s—credibility.

When they do, the stories can ruin lives. Sometimes it is the subject of a piece that is treated unfairly. When somebody planted a bomb at the 1996 Olympics in Atlanta, the Atlanta Journal-Constitution, CNN, and NBC labeled a security guard the prime suspect, based on leaks from anonymous law enforcement officers. The suspicions turned out to be wrong and prematurely reported, but the rumors dogged the guard for years. In 1999, an explosive report detailing the theft of atomic weapons secrets by China in the New York Times relied on anonymous sources to whip up a hysteria that quickly centered on Wen Ho Lee, a Chinese-American scientist at Los Alamos. After being jailed and interrogated, Lee was exonerated of any substantial wrongdoing.

Other times, it’s reporters who take the fall. Jayson Blair, a young reporter for the New York Times, left “a long trail of deception,” in the paper’s words, fabricating stories and details from whole cloth. In 2003, his reporting on the case of snipers in the Washington, DC area relied on unnamed sources within law enforcement, who turned out not to exist. More recently, CNN published a story on potential ties between Trump administration officials and a Russian investment fund, based on a single anonymous source. The flimsiness of that story ended with two editors and a reporter being forced to resign.

It is the belief of NYU’s journalism faculty that anonymous sourcing should only be used when absolutely necessary, and with great caution, and many articles that rely principally on such sources should not run at all.

Daniel Okrent, then-public editor of the New York Times, summed up the issue of “coddling sources” in his post-mortem of the paper’s Iraq coverage in 2004: