Inference: why comprehension is not just about vocabulary and knowledge 

Professor Jane Oakhill explains how to teach comprehension skills, such as inference, in schools 

Jon Severs

Inference: how to teach a key skill for comprehension

The way many schools approach comprehension can be, according to Professor Jane Oakhill, a little too simplistic. 

The professor of experimental psychology at the University of Sussex, and leading expert on the teaching of comprehension and inference, explains on the latest episode of Tes Podagogy (below) that, for starters, comprehension is often viewed as a step on from decoding, when actually it should be nothing of the sort.

“We cannot see comprehension as something that comes after decoding,” she says. “Schools should be teaching comprehension and language skills from the start. The danger of not doing this is that children can come to believe that reading is about decoding, about getting the words right. We know that this is something poor comprehenders do believe.”

She explains that teaching comprehension skills does not require a child to be able to read and so can be started at the earliest opportunity. 

Inference training

“So much can be done with oral language,” she reveals. “Something I have said to a lot of teachers is that children do not have to be reading in order to learn how to understand written text.

"Some of the best instruction in comprehension I have seen is where teachers have encouraged lots of discussion, addressing specific questions in groups and then feeding back to the class. Lots and lots of talk. This is so important and it is very often missed out.”

She says another key part of teaching comprehension can also become distorted in schools: ensuring pupils have a firm base of vocabulary.

A child having a broad vocabulary is key to comprehension, she explains, but it is an issue often tackled via the teaching of word lists. Doing this alone, she says, does not work. 

Quick read: The importance of inference 

Quick listen: Why oral vocabulary is so important at primary

Want more articles like this? Join our Tes Teaching and Learning Facebook group

“The evidence is that most of a child’s vocabulary comes from reading, not through being explicitly taught words,” she explains. “In fact, the meta-analyses of vocabulary instruction are not very promising – it is really quite hard to teach older children new vocabulary effectively. It’s possible, but the transfer effects are not great. In general, it is agreed that explicit teaching of vocabulary is not the way forwards.

“Instead, we should be teaching children to build their own vocabulary. Skilled readers form a general impression of what a word means from the context, and refine that meaning from the context. This is what we need to teach children to do, to try and work out what is going on.

“For example, if the teacher knows that children do not know the meaning of a word, rather than just telling them it would be much more effective to have a discussion about what the definition could be – has anyone got any ideas? Does it fit with the text? If not, why not? It has to be an active, not a passive, process.”

Knowledge comprehension

She also explains that we often believe children to have mastered vocabulary, but actually the depth of understanding they have is not sufficient for comprehension.

“Breadth of vocabulary is one measure – you may know a lot of words but at a very superficial level,” says Oakhill. “They can match a word to a picture, for example. There is increasing evidence that depth of vocabulary is important – a rich meaning network of associations around a word. This is really important for inference skills.”


That "meaning network" is often tackled via explicit knowledge teaching, but Oakhill has similar concerns about relying on that method as she has with explicit vocabulary teaching. 

“I think that it is going to be impossible to teach children this huge bank of knowledge they need to know to tackle texts they may come across,” she explains. “Clearly, knowledge is important in comprehension. It is a really good idea for teachers to teach children to activate and use background knowledge in their reading – and some children don’t do that naturally. 

“But what happens when they do not have any knowledge of a topic? You could give them the rudiments, but are you going to do that every time for the rest of school, and then what about when they leave school? What happens when the teacher is not there with the knowledge base in a package for them? And we have evidence that, even if you give children the knowledge, some cannot make inferences still.

"Like with vocabulary, we need to get children actively engaged with the process.”

Mixed approaches

This is not to say teachers that should avoid knowledge and vocabulary instruction, rather they should be looking at the skills-based training of pupils, too. 

“We should be looking at how the skills and strategies for comprehension can support vocabulary and knowledge, and vice versa: how knowledge and vocabulary can be used to support the skills,” she explains. 

She says there are some core skills we should focus on, the most obvious being inference. 

“The skills that seem to be important in comprehension are, firstly, inference abilities, so linking up information in the text – like ‘he’ refers to someone who is male in the text – and more global inferences, where the child is putting ideas together in the text and relating it to background knowledge,” she explains. “There is a lot of evidence showing that training in inference skills is effective in improving comprehension, but it does have limited transfer to standardised assessments of comprehension.

“Another area is metacognitive understanding, the ability to reflect on your own understanding, knowing you don’t get it and knowing what to do if you don’t get it.  

"And finally, using text structure, such as subheadings, as these give you so many clues and pointers as to where the text is going.”

Spotting the signs

On the podcast, she also discusses the role of motivation, how to teach ‘plausible’ inference and why we need to get better at spotting poor comprehenders. 

“Most of the children who are identified as poor comprehenders later on are perfectly good at holding everyday conversations, they are good communicators and good decoders, it is the text they are not getting,” she explains.

“They are getting missed because they are very diligent, they get on with their work, they are good decoders, good conversationalists, and so it might not be until late primary when there is more pressure on reading to learn that these children who are having difficulty become apparent. We really need to identify them much much earlier.”

You can listen to the podcast on the player above, or via your podcast platform by typing in ‘Tes – the education podcast'


Register to continue reading for free

It only takes a moment and you'll get access to more news, plus courses, jobs and teaching resources tailored to you

Jon Severs

Jon Severs is editor of Tes

Find me on Twitter @jon_severs

Latest stories