Scientific Studies Explain the Best Ways to Talk to Children
For most adults, talking to kids can be daunting. And not just in the "watch what you say" way but in the "they won't say anything to me" way. Luckily, science has stepped in to find the best methods for getting actual responses from kids.
There are a lot of contradictory bits of "common sense" out there about kids. They lie all the time. They're innocents who don't know how to lie. They're only saying something because their parents told them to say it. They always say exactly what their parents don't want them to say. They say the darnest things. They say nothing.
Really, what the research says is that all of these are true. What kind of response you get is dependent on how you act, though, not on the kid.
The vast majority of information on speaking with children comes from research into the best practices for forensic interviews. Forensic interviews are interviews where the goal is to get a child to tell you, truthfully, everything he or she saw or happened to them. Most of the time, it's done when a child is the witness to a crime or the victim of one. For obvious reasons, these interviewers are always looking to minimize the chance that their questions will bias the answers children give, and are always trying to get children to do as much of the talking as possible.
Currently, most of the current understanding of talking to children is incorporated into the National Institute of Child Health and Human Development (NICHD) Protocol and the Ten Step Interview by Thomas D. Lyon, J.D., Ph.D.
Thankfully, most of us aren't going to be talking to children for this purpose. We just want to be able to have a conversation with a child.
Narratives, Open-ended Questions, and Invitations
If you're trying to have a conversation with a child, these three things are going to be the most important. The NICHD Protocol and the Ten Step Interview both include what rapport building/narrative practice. That's basically what any conversation with a kid should look like. Since, in that situation, the practice is to get the child comfortable before asking about something sensitive. But you are just asking about some event and let the child tell you all about it.
However, when a child is not responsive, a lot of us have the instinct to "help" them by giving them "easier" questions. So the invitation "Tell me about your dog" becomes a series of close-ended questions where the adult ends up doing most of the talking. Like:
"Is that your dog?"
"What's his name?"
"Is he brown?"
And the child is just saying "Yes", "Buster", and "No".
Open-ended questions are questions that can't be answered by "yes" or "no" or a single word. In adult conversations, we use closed questions all the time without really knowing that they're closed. For example, most adults hear "Is that your dog?" as an invitation to tell you all about their dog. Kids only see a yes/no question, and will stop talking after giving an answer.
What you want are open-ended invitations that get narrative answers from children. Invitations sound like like "Tell me everything about X" or ""You said X. Tell me everything about that". A 2007 study, looked at the interviews of 52 youths who were victims of sexual abuse. The details given by the children were more likely to be confirmed by another source when they were given in response to invitations.
Making sure the invitations are open ended is equally important. In a 1997 study, Sternberg et al. found that children 9 and older, produced around 50 details when given yes-no prompts and 140 details when narrative practice was used. Open-ended questions mean that children will tell you more. Yet another study not only found that "free-recall" (i.e. open-ended) questions produced the most response from children, but that closed-ended questions produced more details only as children get older.
When the children are older, they've learned not to take close-ended questions as literally as younger children do. They are closer to adults, who understand close-ended questions as invitations to tell a narrative. But for most children, close-ended questions will elicit the shortest answer possible. More than that, close-ended questions may even get a child to tell you things they know aren't true. They feel pressured to answer close-ended questions.
For example, a study by Amanda H. Waterman, Mark Blades, and Christopher Spencer titled "Do children try to answer nonsensical questions?" In this study, 73 children aged between 5-8 were asked 3 sensible open-ended questions (ex: "What do birds eat?"), 3 sensible close-ended questions (ex: "Is summer hotter than winter?"), 3 nonsensical open-ended questions (ex: "Where do circles live?"), 3 nonsensical close-ended questions (ex: "Is a box louder than a knee?"), and two "scrambled" questions (ex: "Than is louder thunder whisper a?").
All the children were told that they could say they didn't know the answer and, when asked three weeks later, nearly all of the children correctly identified the questions as making sense or being "silly". Almost all of the children answered the sensible questions, both open and closed, correctly. And 90% of the children answered the open-ended nonsensical questions correctly, saying they didn't know. However, 72% of children tried to answer the nonsensical close-ended questions, even though they later said that the questions were silly.
Now, as much fun as it is to ask children nonsense questions and watch their faces as they try to answer them, if the goal is have a conversation, stay away from close-ended questions.
This is something that only really comes up in the forensic interview context, but the implications are important for anyone talking to children. The first is letting a child know that he or she can answer "I don't know" at any point. A 1999 study asked 157 children aged 9-13 a number of misleading and neutral questions, both open- and closed-ended, about their trip to a science museum. In the first part of the study, the if the children attempted to answer questions they couldn't know the answer to, the interviewer reiterated that they could say "I don't know" and asked again. This did decrease the answers to the misleading questions but also decreased answers to the non-misleading ones.
This implied that restating the "I don't know instruction" throughout the interview made the children believe that this was a correct answer. Basically, children are incredibly susceptible to the cues you give when you talk to them. In the second part of the study, the "I don't know" instruction was expanded to also encourage giving correct answers, which fixed the problem. For everyday talking to children, this indicates that you should watch your own responses. You may not think you're doing it, but smiling at a certain answer or giving a certain response may make the kid repeat the answer, in hopes of pleasing you.
(Note: This study also reinforced the open-ended questions thing: even when told they could say "I don't know", children had trouble not answering the close-ended misleading questions)
The other interesting one is a 2008 study into truth induction. In this study, 99 boys and 99 girls aged 4 to 7 spoke with an interviewer who then left the room. Another adult entered and did one of four things: 1) play with coin and Lego house; 2) play with coin and Lego house, but the adult encouraged kid to keep the play a secret; 3) only play with coin; and 4) play with coin, but child encouraged to say they played with the house. Then, the interviewer returned and did one of three things: interviewed the child normally, reassured the child that they could tell the interviewer anything and they wouldn't get in trouble, or ask the child to promise to tell the truth. Reassurance did not have a positive effect, but the oath did increase the number of children to tell truth, regardless of what had happened in when they played with the other adult. This implies that asking a child to tell you the truth will greatly increase the chance that that's what you get.
Don't Try to Get Time and/or Numbers
Always remember that time and numbers are really difficult for children. One study by Wandry et al., asked 167 children aged 6-10 years to name their age, birthday, and current month. By age 10, almost all of the children were answering the questions correctly. Additionally, all of these children were in foster care. Half were asked about their last move and half were asked about their last court date, which would appear to be events that would stick in a child's mind. However, even only 56% of the older children could correctly answer questions about their most recent move and only 52.78% of them could correctly answer the questions about their last court date. That's barely better than chance.
Yet another study staged an event around Halloween for 86 4 to 13-year old children. Three months later, the children were asked about the event. Fewer than 20% remembered that it had taken place close to Halloween, and they were no better than chance when asked if it was before or after Halloween.
When talking with kids, stay away asking about time. Chances are that they won't be able to tell you when something happened.
In terms of numbers, the same Wandry et al. study above also looked at how accurate children were in answering questions about how many times they've moved or been to court. Very, very few of the children could correctly give the number. And only about two-thirds of the children could correctly answer the more leading "One time or more than one time?" version of that question.
Other Things You Can Do To Make a Child Talk (And Things Not To Do!)
Here are a few other things that studies have shown encourage a child to speak:
- Using the child's name (Hershkowitz, 2009)
- "Back-channel" facilitators like "Uh-huh" and "Oh" (Cautilli et al., 2005)
And a few things not to d
- Unclear invitations like just saying "Tell me more" (Hershkowitz, 2011) or even "Tell me more about that" (Walker, 1993)
- Invitations as questions. This is a huge one for those of us from California. The invitations "You said X. Tell me more about X" can become "You said X?" with the right inflection. That turns it into a yes/no question and will pretty much stop a kid's narrative. (Evans & Roberts, 2009; Evans et al., 2010)
Thanks to Thomas D. Lyon, J.D., Ph.D., whose seminar on child interviewing inspired this article.