Category Archives: methodology

Interviewing children without inducing the answers

When we interview children for research purposes we usually face two typical dangers:

–          Suggestibility: sometimes researchers use biased questions that contaminate children’s beliefs or memories. Sometimes children just want to give interviewers the reply they feel is expected or “socially desirable”. It’s easy to induce a given response in a 4- or 5-year-old, and if this happens… then research results are worthless.

–          Unreliable memories: Not only are young children very suggestible, they are also not reliable when recalling events that just took place. Thus, if we show them a video clip and then ask them some questions about it, it’s important that we make sure that children understood what they saw and can retain the events in their memory.

 

To sum up: it is essential, when we are designing our research interviews, to avoid any tricky questions or stimuli that might interfere with children’s spontaneous thinking (the latter being what we are interested in). I’ve just read three papers that supply interesting and relevant findings we should keep in mind when designing appropriate research interviews with young children.

1) Roebers & Schneider (2005) found that the better the child’s language abilities the less suggestible the child is. Language development seems to be key (especially language fluency and comprehension) rather than other general domain variables such as executive function or working memory. Investigative interviews are language dependent; language abilities play a major role for explaining differences in suggestibility. This finding, however, comes with an interesting caveat: it is also easier to purposely disrupt children’s memories when they have good language skills. The reason for this is that children with better language skills process (false) verbal information provided by the researchers more efficiently, and later they have trouble distinguishing between original and suggested information. Individuals with better language skills encode, store, and thus remember the contents of the misleading questions better than do individuals with poorer language skills. Language can work either way.

2) Peterson, Dowden, & Tobin (1999) investigated the influence of question format on 3 to 5 year old children. They found that when researchers frame their questions with a yes/no format (e.g., “did the woman take the man’s hat?”) many preschoolers tend to reply “yes”. This is so even in cases when they don’t know the answer or when they have been presented information that requires a “no” answer. By way of contrast, when children are asked the equivalent wh- question (e.g., “what did the woman take?”) children give more accurate answers and the percentage of children who answer “I don’t know” increases. The authors conclude that there are dangers inherent in yes-no questions: answers may be influenced by response biases or other factors besides how veridical the underlying proposition is. Children seldom say “I don’t know” when they are uncertain or do not know the correct answer. Specific wh- questions seem to be less problematic.

3) Mellor & Moore (2014) investigated elementary school children’s ability to use Likert scales during research interviews. In my opinion, their work has several important methodological flaws. Mainly, the questions are too complicated and children’s failures to use the scales adequately reflects, in my opinion, more the inherent difficulty of the problems posed to children rather than the (un-)reliability of Likert scales. There are a few interesting comments in the paper, anyhow: a) children tend to respond to Likert scales with a left-bias (that is, they seem to pick the first item in the scale more often than the rest); b)  5-point scales yield similar results to 3-point scales (they don’t seem more difficult to understand for elementary school children); c) word-scales (eg: very good-kind of good-more or less-kind of bad-very bad) produce more reliable results than number scales (5-4-3-2-1).

Altogether, three interesting and relevant papers. I’ll keep these findings in mind when designing my own research interview.

Mellor, D., & Moore, K. A. (2014). The use of likert scales with children. Journal of Pediatric Psychology, 39(3), 369–379.

Peterson, C., Dowden, C., & Tobin, J. (1999). Interviewing preschoolers: Comparisons of yes/no and wh- questions. Law and Human Behavior, 23(5), 539–555.

Roebers, C. M., & Schneider, W. (2005). Individual differences in young children’s suggestibility: Relations to event memory, language abilities, working memory, and executive functioning. Cognitive Development, 20(3), 427–447.