We may treasure honesty, but the research is clear. Most classic strategies to promote truthfulness just encourage kids to be better liars.
Ashley and I went to Montreal to visit the lab and operations of Dr. Victoria Talwar, one of the world’s leading experts on children’s lying behavior. Talwar is raven-haired and youthful, with an unusual accent—the combined result of Irish and Indian family ancestry, a British upbringing, and stints in American, Scottish, and Quebeçois academia. Her lab is in a Gothic Revival limestone mansion, overlooking the main campus of McGill University.
Almost immediately, Talwar recruited us for one of her ongoing experiments. She threw us in a small room with two of her students, Simone Muir and Sarah-Jane Renaud, who showed Ashley and me eight videos of children telling a story about a time they were bullied. Our role was to determine which kids were telling the truth and which had made their story up, as well as to rate how confident we were that our determination was correct.
The children ranged in age from seven to eleven years old. Each video segment began with an offscreen adult asking the child a leading question to get the story started, such as, “So tell me what happened when you went to Burger King?” In response, the child told her story over the next two and a half minutes, with the occasional gentle prod for details by the adult who was interviewing her. Those two-plus minutes were an extensive length of time for the child, offering plenty of chances to include contradictory details or hints that might give away her lie.
This format was crafted to simulate the conditions of children testifying in court cases, which is where the modern science of kids’ lies began. Over 100,000 children testify in American courts every year, usually in custody disputes and abuse cases.
In those cases, children are frequently coached by someone to shape their story, so the children in Talwar’s experiment were also coached, briefly, by their parents the night before. To prepare the videotape, each child rehearsed a true story and a fabricated tale, and told both stories to the interviewer on camera. The interviewer herself did not know which story was true. Then, one of the child’s stories was included in the videotapes of eight. The stories chosen for the tapes were not picked because the child did an especially great job of lying. They were merely picked at random.
The adorable little girl with the Burger King story told how she was teased by a boy for being Chinese, and how he threw some French fries in her hair. I froze—would a total stranger throw fries in a girl’s hair? She looked so young, and yet the story came out in full, complete—rehearsed? Just guessing, I marked this as a fabrication, but noted my confidence was nil. My confidence didn’t improve with the next two children’s stories.
“This is hard,” I murmured, surprised that I didn’t have the answers immediately. I pushed myself closer to the video monitor and cranked the volume up as loud as it could go.
Another girl told of being teased and left out of her group of friends after she scored 100 on a math test. She told her story with scant details and needed a lot of prodding; to me, that seemed genuine, childlike.
After the test, Ashley and I were scored. To my dismay, I got only four right. Ashley got only three correct.
Our results were not unusual. Talwar has run hundreds of people through this test, and on the whole, their results are no better than chance. People simply cannot tell when kids are lying. Their scores also tend to reveal some biases. They believe girls are telling the truth more than boys, when in fact boys do not lie more often. They believe younger kids are more prone to lying, whereas the opposite is true. And they believe introverts are less trustworthy, when introverts actually lie less often, lacking the social skills to pull off a lie.
There are many lie-detection systems created from the patterns in verbal and nonverbal behavior in adult lies, but these provide only small statistical advantages. Voice pitch, pupil dilation, eye tracking, lack of sensory details, and chronological storytelling are some indication of lying in adults. However, when accounting for the wild standard deviation of these behaviors in kids, those higher-than-average indicators become not much more reliable than flipping a coin.
Thus, police officers score worse than chance—at about 45%. Customs officers are trained to interview children during immigration processing and instantly determine if a child has been taken from his parents. Yet they, too, only score at chance on Talwar’s test.
Talwar’s students Muir and Renaud have run several versions of the experiment with both parents and teachers. “The teachers will score above chance—60%—but they get really upset if they didn’t get 100%,” said Muir. “They insist they’d do better with their own students.”
Similarly, the parent’s first defense against his child’s tendency to lie is, “Well, I can tell when they’re lying.” Talwar’s proven that to be a myth.
One might object that these bullying videotapes aren’t like real lies, invented under pressure. They were coached, and the kid wasn’t trying to get away with anything.
But Talwar has a variety of experiments where she tempts children to cheat in a game, which puts them in a position to offer real lies about their cheating. She videotapes these, too, and when she shows those videotapes to the child’s own parent—and asks, “Is your child telling the truth?”—the parents score only slightly better than chance.
They don’t take it well, either. When Renaud’s on the telephone with parents to schedule the experiments, “They all believe that their kids aren’t going to lie.” Talwar explained that a number of parents come to her lab really wanting to use their kids’ performance to prove to a verified expert what a terrific parent they are.
The truth bias is a painful one to overcome.
The next day, we saw that in action.