Twenty questions about your family, and what the research behind them really found

The twentieth question is the odd one out. After nineteen sensible items about where your parents grew up and how they met, it asks: "Do you know about a relative whose face 'froze' in a grumpy position because he or she did not smile enough?"

Nobody's face has ever frozen. That's why it's there. You can only answer question 20 if someone in your family told you the story, probably more than once and probably while somebody else at the table rolled their eyes. It's the one item on the list you can't look up.

The list is the "Do You Know?" scale, written by the Emory University psychologists Marshall Duke and Robyn Fivush. Duke published all twenty questions himself in 2013. They're worth reading in full. So is what the research behind them found, which is more modest and, I think, more useful than the version that usually gets shared.

Where the list came from

Duke told the writer Bruce Feiler that the project began in the mid-1990s, when he was asked to help study myth and ritual in American families. His wife Sara, a psychologist who worked with children with learning disabilities, had noticed something about her students: "The ones who know a lot about their families tend to do better when they face challenges." (Feiler's 2013 New York Times column, republished on his own site.)

So they wrote twenty yes-or-no questions. Duke has said they were chosen because they "test knowledge of things that children could not possibly have learned first hand." You weren't there when your parents met. You weren't there when your grandmother left the town she grew up in. If you know about it, someone told you.

A few of them, to give you the flavour:

• Do you know how your parents met?
• Do you know the source of your name?
• Do you know which person in the family you act most like?
• Do you know some of the lessons that your parents learned from good or bad experiences?
• Do you know some of the jobs that your parents had when they were young?

What they found

In the version Feiler reported, the scale "turned out to be the best single predictor of children's emotional health and happiness," and children who knew more had a stronger sense of control over their lives. There's also a striking coda. The team had tested a group of families in 2001; two months later came September 11. When they went back, Duke said, "the ones who knew more about their families proved to be more resilient."

The published study people usually cite is smaller and more careful. The Emory team gave the scale to 66 teenagers aged 14 to 16, and those who knew more about their family's history reported higher emotional well-being and a stronger sense of identity, even after accounting for how well their families functioned overall (Emory's write-up, via Phys.org). The same write-up says plainly that this was the only study so far to use the scale, and that more research was needed before drawing firm conclusions.

What the list doesn't show

Sit with those numbers for a second: 66 teenagers, middle-class, from two-parent families, in the United States. And it's a correlation. A teenager who can answer twenty questions about their grandparents probably lives in a family that talks a lot and has time to, and either of those could be doing the work. I'd be wary of anyone who tells you the questions build resilience.

Duke would be too. In the same piece where he published the list, he wrote: "Simply knowing the answers to questions will not produce the good outcomes described above." What mattered, he argued, was how children came to know it, and in the families they studied that happened "during family dinners, family vacations, family holidays, and the like." The quiz measures whether the telling happened. Drilling a child on the answers would miss the point entirely.

Fivush's later work makes the picture messier, in a way I find reassuring. In a 2011 study, 65 adolescents told stories about their parents' childhoods. Teens whose stories took the parent's point of view, and drew connections between the parent's life and their own, reported higher well-being. But the link only held for girls telling stories about their mothers. Nobody seems quite sure why. It's the kind of result that doesn't fit on a poster, which is probably a sign it's real.

The shape of the story

One more idea from Duke has stuck with people. He described three shapes a family's history can take. In the ascending one, we came here with nothing and worked our way up. In the descending one, we used to have it all and lost it. The third, which Duke called the most healthful, oscillates: we've had ups and downs, and we stuck together through them.

As far as I can find, that's Duke's clinical judgement rather than a measured result, so hold it lightly. It still rings true to me. A child who has only heard the triumphant version has nothing to set a bad year against. A child who knows that their grandfather lost a job, and that the family got through it, has a precedent.

The table moved

The research pictured children at a dinner table, picking up stories sideways while the adults talked. A lot of the people reading this live in a different city from the parents who hold those stories, sometimes a different country. The table, for many families, is now a WhatsApp group, which is lovely for photos of dinner and hopeless at keeping anything you can find again in six months.

The questions also show where they were written. Question 15 asks about your family's "national background (such as English, German, Russian, etc)." In plenty of families the better question is narrower: which village, which district, what language your grandmother's mother spoke at home, why the family moved and in what year. Rewrite the list for your own family. Duke himself called it "a representative sample," not a fixed test.

And it works on adults. Try it on yourself before you try it on anyone else. My guess is most of us do fine on the first few and go quiet around the illnesses and the jobs.

One thing to do this week

Go through the twenty yourself and mark the ones you can't answer. Pick one. Rewrite it in your family's terms, and ask the person who'd know.

Ask one question, not twenty. A list of twenty reads like a form, and nobody tells a story to a form. Ask in a way that lets them talk, a call or a voice note, because the answer to "how did you two meet?" sounds different from the version someone types out with one thumb. An invented example of a rewritten question, for a Bengali family: "Ma, what did Dida cook when someone in the house was ill?" It's nowhere on Duke's list, and it'll probably get you a recipe and a story in one go.

When the answers don't match, leave them be. If your mother's account of how she met your father isn't quite your aunt's, you haven't found an error. You've found the oscillating version.

Conneczen is built around that single question. You send one ask as a link, over WhatsApp or anything else; the person opens it and records the answer on their own phone, without making an account; it comes back in their voice, with their name on it, in a scrapbook your family can share. If that sounds useful, you can start one here.