Why are some children smarter than others?

Being smart refers to the mental skills that help us to adapt successfully to the environment. We can think of the world around us as a series of challenges: objects that quickly move toward us; people who talk to us; things that we want to do or possess. Each of these challenges involves cognitive ability, for example the ability to navigate, to speak, and to plan. Some people meet these challenges better than others, because their abilities to learn and adapt are better. In other words, being smart means figuring out what to do in a situation, catching on, making sense of things, and learning from experience.

Some children have an easier time to learn than others. Yet, all children learn — some just need more time and support. A straightforward consequence of children’s differences in the ability to learn are their differences in school grades. Some children come top of their class and quickly master the curriculum, while others need extra tutoring and have to practice harder and then they may still get lower grades.

Of course, school grades aren’t everything and children can be smart in many other ways that are not limited to the school environment. For example, some kids are amazing athletes, others beautiful singers, and again others know the best hide-outs in the neighbourhood. Still, school grades are very important: In the UK, like in most other countries, school grades regulate the access to further education. For example, teenagers’ GCSE grades at the end of compulsory education will influence their decision to go on to A-levels next, which in turn will affect if and which university they will apply to.

The 30 million word gap

In the 1990s, Betty Hart and Todd Risley published a seminal book on their efforts to understand the origin of children’s differences in language development. Both researchers had taught children in kindergarten and primary school settings, and they had come to realize that children differed substantially in verbal ability. These differences in verbal ability were already evident before children had started kindergarten or school, which suggested that they stemmed from children’s early life experiences in their family homes.

To test this idea, Hart and Risley started a longitudinal study that followed 42 families with infant children over the course of two and a half years. Each month, research assistants visited the families and observed them for one hour. To do so, the research assistant would follow the family child with a microphone and tape recorder in hand, trying to document everything the child heard and said over the course of an hour. Back at the lab, the research assistant then transcribed the recording and prepared the data for analysis.

Later Hart and Risley extrapolated the data from the hour-long recordings to match 24-hour days and thus, estimated how many words each child heard and spoke per day. The made two key discoveries.

First, how many words children heard over the course of a day differed greatly across families. Children heard pretty much the same types and kinds of words, like ‘sleep’ and ‘eat’ and ‘nice’ but some kids heard them much more often. In fact, it was estimated some children heard 30 million fewer words than others by the time they are 4 years old.

The second finding was that children who heard more words at home also developed better verbal ability themselves, which in turn was linked to better school performance. In other words, children from families that spoke more developed greater language skills that then helped them do better at school.


At the Hungry Mind Lab, we sought to replicate and extend Hart and Risley’s work. Specifically, we wanted to test a larger sample, from Britain, and observe the children unobtrusively for longer than an hour — without a research assistant running after them with a tape recorder in hand.

With funding from the Wellcome Trust, we recruited overall 120 families from South East London, who had young children aged 24 to 48 months. Each family completed an extensive background survey online, before they had a box with the study materials hand- delivered to them.

The box included LENATM digital language recorders, about the size of a floppy disk, that children wore inserted in designated chest pockets in their t-shirts. Parents also received a testing booklet to complete together with their child to assess the child’s cognitive development.

Parents completed 3 full days of digital recordings with their children, as well as the cognitive testing booklet over the course of the next month. On recording days, the digital language recorders documented everything that the child heard or said in a six-foot radius. Once the recordings were done, a research assistant visited the families again to pick up the study box with its materials.

The 30 million word gap, revisited

We finished collecting all data in March 2017 and have been analysing the data since. First, we found that families differed vastlyed for their children. On average, children heard about 18,000 words per day but the range was enormous. In one family, the child heard 3,200 on a day, while in another the child heard 45,500 words. That’s a 14-fold difference! Our full paper on these results is here.

We also found that how many words a family spoke differed from day to day and hour to hour. It seems obvious that families don’t always speak the same amounts. They are likely to chat more over meals and less during nap and play time, when children are by themselves. But across entire days, we had expected that chattier families always speak more than quieter families.

Yet, our data showed that about half of the differences in the number of words that children hear differs across days within their families. In other words, families change how much they talk among each other from day to day.

This finding is important because it challenges the notion of a 30 million word gap. It seems that children’s early life language experiences fluctuate greatly over time, which makes it difficult to put a precise estimate on how many words a child will hear.

In the summer of 2019, when the children in our study were 4 to 8 years old (time flies!), we followed them up and collected data on their school performance and development. We are now testing if children’s early life language experiences predict their developmental outcomes in primary school. Our project preregistration and our preliminary findings are here.