A journalist’s take on reporting with unreliable narrators

Part way through the first draft of a major feature, I realized something was off. I was investigating the post-pandemic future of gambling in Nova Scotia for The Coast and looking for an avid gambler to explain how COVID-19 impacted their wagering. “John” responded to my request in a local news Facebook group.

In some ways, John was a journalist’s dream. He was eager to talk and open about his story, even willing to do an pandemic-era, in-person interview in November 2021, during a narrow window I was in Halifax. He offered up details that weren’t necessarily flattering — though I’m not sure he viewed them that way. Nights napping on a casino couch to get a 6:00 a.m. bus because he gambled away his cab fare, for instance, was part of a hero’s journey for John, even if it doesn’t appear that way to an onlooker.

I’d interviewed sources about complex topics before — including people recovering from gambling addictions. But after the first interview, I realized John’s case was more challenging. Not merely because the other sources had parts of their stories backed up by articles and legal documents, but because he had a fairly rosy view on gambling and how it impacted his life. When someone believes they’ve been enriched by a habit, as opposed to damaged, their experience and the stories they want to tell about it changes. Many of his experiences were not easy to verify or quantify. For example, John told me his grandmother’s boyfriend played in a pool tournament on a Las Vegas trip referenced in the story. I feel confident that I tracked down the exact event, but John couldn’t recall enough specifics (like a narrow-enough time window) for me to feel comfortable reporting it by name and expanding this line of detail with more colour. Other stories had significant plot holes or contradictions. I didn’t feel he was being intentionally deceptive (and still don’t), but I started worrying about the precision with which he portrayed himself. I wanted to present John accurately and factually, while not glamourizing gambling or relying too much on difficult-to-prove facts or his best guesses. I worried that if I wasn’t careful with my framing, his portrayal could veer into fiction: Jimmy from the Janet Cooke story, though I’d met John in the flesh and had his words on tape.

I came to realize the challenge with John was that he was an unreliable narrator; a literary term for a storyteller whose tales are compromised and require some suspension of disbelief. The concept dates back to a 1961 book on fiction theory from U.S. literary critic Wayne C. Booth. In journalism, there’s little talk about them outside documentaries; unreliable narrators are mostly analyzed in fiction or media like scripted films and television. (Think Gone Girl’s Amy and Nick Dunne or American Psycho’s Patrick Bateman.) As journalists, we put a lot of trust into those we interview, and sometimes we could benefit from more skepticism and rigour. Journalists try to approach the craft with fundamental ideals of accuracy, fact-checking and objectivity. But an unreliable narrator presents obstacles to some of these basic tenets of the trade.

Though the term unreliable narrator sounds malicious, it isn’t. While some may attempt to mislead or obscure in bad faith or for fun, there are many reasons an interviewee may do so. Some narrators are unreliable due to a lack of age or experience. There are some people who may provide unreliable narration while experiencing a form of mental illness affecting their view on reality. This requires careful consideration of other factors, such as the ethics of interviewing the person and their ability to provide informed consent. But for many unreliable narrators journalists may encounter, this isn’t the case.

Others use embellishment to reimagine their own disappointments or to project themselves more grandly. Consider the Caliphate podcast scandal, which resulted in the New York Times retracting its core story after it came to light the outlet had been duped by a key source.

“We fell in love with the fact that we had gotten a member of ISIS who would describe his life in the caliphate and would describe his crimes,” Dean Baquet, then executive editor of the Times, told NPR in 2020. “I think we were so in love with it that when we saw evidence that maybe he was a fabulist, when we saw evidence that he was making some of it up, we didn’t listen hard enough.”

As I grappled with John’s interview, I reflected on lessons like those learned at the Times and through my own investigation.

It’s critical to admit that journalists are fallible and need to be vigilant about narratives that fit our purpose or seem compelling. Despite myself, I felt this with John. He made winning money sound easy, and I wondered, momentarily, if I could use his secrets to stretch my savings. But when I asked about his secret and it became clear he didn’t have one, the revelation had a sobering effect, like taking a cold shower. It’s important to acknowledge that the possibility of being taken by a story or source is a normal feeling — if we can’t, we’re bordering on unreliable ourselves. Then we must keep that in mind when reporting.

It’s hard to realize someone’s an unreliable narrator until they’re being interviewed. I was helped immensely by a close review of the interview transcript, affirming for me how important it is to record sources. Accuracy, context, intention and the ability to parse all a source’s words, not merely conversation fragments, is critical. When I went through the transcript I thought about the range of what I trust from most interviewees as a best practice, and used that context to look at his stories. We trust interview subjects who offer opinions, memories or describe personal experiences. We don’t expect perfection from memory, and journalism allows for that with hedging rather than declarations. (For example, John is pretty sure he lost money overall in November 2021 in my piece; the most he recalls losing is $10,000.)

I was relieved to realize that a lot of what he talked about was information I was comfortable trusting from most sources. Broadly speaking, the conversation wasn’t concerning. Only some small details from it elicited heightened scrutiny. This makes sense, because certain facts or omissions feel larger than others, though they might only represent a few seconds or minutes of conversation. We can’t let this cause us to throw the baby out with the bathwater.

One way to patch some plot holes is using documents or other information from web searches. I didn’t have much from John, aside from Facebook. But scrolling through his profile I was able to confirm different things about his life, like that he was telling the truth about having children (including some of their identifying characteristics) and the frequency of his trips to the casino. It was clear he didn’t invent them as part of a tale to tell a reporter — there was too much history to back them up. While at one point I’d hoped to include some details from the spreadsheets he says he kept pre-pandemic to track gambling outcomes, after our second interview we had a conversation that changed things. John told me he was glad I was writing a positive story about gambling — something I can assure you was not how I presented it. After I dispatched this notion, he stopped replying to me.

This, for instance, is why I avoided speculating too closely on his wins or losses over the years. Elizabeth Stephen, a Halifax-area counseling therapist I spoke with for the story told me gamblers have a tendency to minimize losses and maximize wins. This led me to realize that John’s ballpark estimates — which might seem like reasonable questions — were unlikely to be as reliable as I’d want them to be. By focusing more on his lived experiences in the piece, and tying them into the broader narrative, I felt comfortable with how John was presented, offering a view into how he saw the gambling industry, the world and himself. Because his story was shared along with expert interviews, well-sourced statistics and other research, it was part of an entity that became nuanced and thought-provoking. John’s views presented alone, with little of this added context — say, in a short single-source profile — would be an entirely different story.

Using John’s own words was helpful. He was an engaging storyteller and I ended up including some of our interviews almost verbatim in the piece, italicized like a monologue. I found this helpful to present his opinions and his contradictions without inserting myself as a judgy narrator. I took inspiration here from the way Michael Schulman wrote about Succession star Jeremy Strong in The New Yorker and how documentarian Lauren Greenfield presented the queen of unreliable narrators, Imelda Marcos, in The Kingmaker. These works made me realize that it was possible to let complex people present their version of the truth, if I knew when to push back. Not only in my questioning in interviews, but in my writing. Greenfield’s documentary offered an example that stuck with me: a scene with Marcos talking about how she remembered no challenges or hardships in her marriage followed immediately with archival footage and interviews with others discussing her husband’s extramarital liaisons, and how she used this as leverage with him to get what she wanted. If Greenfield had merely used the quote unchallenged, as Marcos may have expected, she would have been going against a well-documented history. That’s why it’s important to remember the framing of an unreliable narrator’s story might not match reality. Someone like Marcos is extreme, as she has been crafting grand narratives about her life that go against realities for decades, but reporters work small, one project at a time. If an unreliable narrator is offering lies on a small scale it works to achieve their ends.

Between my initial and followup interviews with John, I decided to ask Stephen, who specializes in counselling people with gambling addictions, about some of my concerns. She made it clear that in her work it was necessary not to believe everything presented by someone with a heavy gambling habit. John saw his gambling through a different lens and his biases would colour what he told me.

Originally, I’d thought of ending the piece with his musings on how he felt an improved economy would make him stop gambling, as it seemed interesting and I could understand (at least to a degree as a fellow millennial) where he was coming from. When I told Stephen, she was immediately skeptical and offered a concise counterargument: if John was really concerned about the economy, he could invest money instead of gambling it away, potentially earning bigger (and in some cases guaranteed) returns. That was an aha moment for me, and I included both that and my earlier realization about John’s lack of a winning secret — those cold shower moments — in my the final draft. I felt readers might find them as insightful as I did to understand John. I also felt that it was better not to conclude with his voice — in a piece where that gesture might add further weight to his words, it felt more rational to end with a short bit of analysis.

Overall, I’m pleased with how I portrayed John in the investigation. It received good reader feedback (including from some of the experts quoted) and was nominated for two Digital Publishing Awards. I’m not going to claim perfection. Making edits in hindsight is 20/20 and journalism is like the management concept of continuous improvement — we should always try to get better. But I learned a lot from asking myself questions about John’s story and my approach to my craft. I know next time I’m faced with an unreliable narrator I’ll feel more at ease with how to approach their story. I won’t panic, and I’ll have a greater understanding of how to incorporate their experiences and views into my reporting without letting them take over the tone or avoid harsh truths. I hope that’s true for you too.