Confidence, or at least the appearance of confidence, is regarded as an important attribute in our society. 

Confident people appear more trustworthy and believable.

We see confident politicians in a more favourable light than those who appear uncertain or hesitant, yet we are also somewhat wary of the ones who come across as overconfident.

But are children as sensitive as adults to these cues? And, consequently, do children see teachers and other adults as credible sources of information? 

Confidence and calibration

In a 1989 study, Chris Moore and co-researchers found that children as young as 4 can distinguish between verbal markers of certainty and uncertainty (“know” versus “think” or “guess”). 

These children also preferred to learn new information from a person who appeared confident and certain in their knowledge.

Well-calibrated people are confident when they are informed or accurate but hesitant when they are uninformed or inaccurate. If, on the other hand, we are confident in the absence of knowledge, we are labelled as miscalibrated.

There have been many studies that have looked at how adults use such cues to establish how credible the information given is, but much fewer looking at how children gauge the credibility of adults. 

Eyewitness accounts

A 2008 study provides some clues as to how this might occur in five- and six-year-olds compared with adults.

Researchers presented their participants with stories involving conflicting statements from two “eyewitnesses” to an event where someone threw a ball through a window. 

The first eyewitness stated that Tyler had thrown the ball while the second said that Kenny had thrown it. The eyewitnesses also provided two further pieces of information; both said it was a sunny day, while the first witness said that they were certain the ball was red and the second witness said that it might have been blue. Initially, both children and adults were more likely to believe the first witness. 

Participants were then told that both witnesses were right about the day being sunny, but both wrong about the colour of the ball (it was white).

Armed with this new information, the adult participants now favoured the second witness, but the children still favoured the first one. In other words, the children appeared to be insensitive to calibration as a cue of credibility.

Does this then mean that young children are always swayed by a person’s confidence? Not necessarily. Other studies have found that five-year-olds often take into account previous accuracy, so they are likely under some circumstances to change their views on the credibility or a person if they provide information that they discover to be inaccurate later.

Credibility

Now a new study by researchers at the University of British Columbia, the University of Montana and Oxford Brookes University has attempted to expand on what is currently known about children’s use of calibration.

Researchers set out to investigate children’s understanding of the relationship between how confident or hesitant a person was and the perceived credibility of both the message and the messenger.

They recruited 662 children aged between 3 and 12 years to take part in three experiments. The researchers wanted to identify from whom the children preferred to learn new information; people who are hesitant when uninformed or confident when uninformed. 

In addition, they investigated whether children could recognise when hesitancy was justified and confidence was unjustified. Finally, they looked at how this might impact children’s perception of how “smart” the messenger was.

In one phase of the study, they presented children with videos to establish that one adult model had a track record of being well-calibrated. In the videos, a researcher presented adult models with four covered boxes that the children were told contained different pictures. 

They then asked the models: “Do you know what’s inside the box?” The first model confidently claimed to know the contents of the box while the second was more hesitant. 

In one version (the informed condition) they showed both models the contents of the box, while in the other (the uninformed condition) neither saw inside the box. This meant that in the first condition, the confident model was well-calibrated (their confidence was justified) while in the second condition the confident model was miscalibrated. 

We already know from previous research that children’s understanding of visual access (looking usually leads to knowing) is evident in children from about 3 to 4 years of age (and perhaps even earlier).

Once calibration was established, the researchers assessed which model the children preferred to learn new information from. Would the children in the uninformed condition recognise that there is no justification for the confident model’s confidence and prefer to learn information from the hesitant model? In addition, which model would they consider smarter?

Decisions and development

The results implied a developmental trajectory. Given a choice to learn from a previously confident model, children in the informed condition (when the models were shown the contents of the box) generally preferred to learn from the model who appeared confident, an outcome consistent with previous findings.

However, the results showed that calibration can also influence children’s learning decisions.

Children were less likely to learn from a confident model in the uninformed condition (when the models weren’t able to see what was in the box), implying that the confidence the model displayed wasn’t justified. They were, however, more likely to learn from the confident model in the informed condition.

But there were also differences related to the age of the children. At about 5, children were significantly less likely to judge the confident model as smarter when miscalibrated (that is, in the uninformed condition) compared with well-calibrated (the informed condition).

It would appear that, with age, children increasingly judge hesitant people as smarter if they are well-calibrated (their hesitancy is justified).

One thing to bear in mind is that the study was conducted using videos that the children were asked to watch. It would be interesting to see if the results can be maintained if the study were to be carried out in a more realistic classroom setting. The authors of the study do stress that more research is needed to obtain a better understanding of how children perceive hesitancy.

According to the study’s lead author, children “don’t fully understand what it means to be hesitant and the inferences they apply to whether a person’s confidence is justified don’t get applied to hesitancy”. 

Adults, for example, recognise that responding hesitantly is justified when you don’t know the answer to something, but this didn’t appear to be the case with the children in the experiments.

The study indicates that teaching and learning isn’t only about what we communicate but also how it is communicated. It also offers some fascinating insights into how children develop the ability to distinguish overconfidence and when being hesitant is justified.

Marc Smith is a chartered psychologist and teacher. He is the author of The Emotional Learner and Psychology in the Classroom (with Jonathan Firth) and tweets @marcxsmith