UPDATE: In light of this post on Timothy Shanahan’s blog, I am persuaded that reading comprehension is a thing, but the advice below still stands. 

Most of the schools I visit are unsurprisingly keen to explore ideas to narrow the gap between their most and least advantaged students. Whilst there are also sorts of complex chains of causation which go some way to explaining why children from wealthier backgrounds outperform their less fortunate peers, one particularly vexed question that I’m frequently asked about is that of reading.

The case I’m making here is that reading comprehension should be more properly thought of as language comprehension. Once word recognition has been mastered (phonological awareness, decoding and sight recognition of familiar words) children’s ability to understand what they read is broadly the same as their ability to understand what they hear. If they can’t understand information communicated verbally the root cause is likely to be due to lack of vocabulary knowledge or the lack of  background knowledge that would help them make sense of the information being communicated.

The most significant barrier to comprehension of written texts is that of reading fluency. Too many students in secondary schools simply cannot decode quickly enough to read effortlessly. Anything that occupies our attention reduces our capacity to think and, if some of our limited cognitive resources are being used to blend graphemes into phonemes, then we’ll have fewer resources with which to consider meaning. The optimal speed of reading for comprehension is somewhere around 250 wpm*; if reading speed falls much below this speed then comprehension quickly tails off. Have a go at reading the text in the video below:

Now ask yourself the following questions:

  1. How far did they climb?
  2. Where did the characters find themselves?
  3. At what point did they first see Pemberley House?
  4. Where were they standing when they first saw the house?
  5. How did the author describe the road?

It’s surprisingly hard. You probably got the first question right easily enough but after that you may well have found yourself guessing. Obviously the constraints are highly artificial but reading like this allows us to glimpse the lived experience of an estimated 20% of children in secondary age students in the UK. If this is in anyway similar to your experience of reading it’s almost certainly not something you’d do for pleasure.

If you read the same text without the constraints, comprehension becomes much more straightforward:

They gradually ascended for half a mile then found themselves at the top of a considerable eminence where the wood ceased and the eye was instantly caught by Pemberley House situated on the opposite side of the valley, into which the road with some abruptness wound.

Barriers now are caused by the possibility that some of the words in the passage are unfamiliar (ascended, eminence) or that the syntax of the sentence is unfamiliar and hard to follow. Now consider the following question: What is the writer suggesting about Pemberley House? The phase is “the eye was instantly caught” suggests there’s something visually striking about Pemberley House, but gives no indication of what this might be. But, if you recognise that Pemberley is the ancestral home of Mr Darcy in Jane Austin’s novel Pride and Prejudice, then you will probably know it’s a stately home and it’s likely to be considered striking because of its size and beauty. Even if you didn’t know this, you might still have assumed that a house with a name rather than a number in the middle of wooded countryside is likely to be fairly swanky. Whether or not you know these things has nothing to do with your reading ability per se.

Why does this matter? Well, if you believe reading comprehension is distinct from language comprehension then you might well be persuaded that it’s worthwhile to spend curriculum time teaching reading comprehension. I’ve written about how this plays out in English lessons here and here. I’ve also written about the promise of following an approach which dispenses with trying to teach comprehension skills and focusses instead on just reading. Essentially, children are more likely to improve their reading comprehension by reading more than by being taught more about inference and analysis because they’ll encounter a wider range of vocabulary in a wider range of contexts beyond their direct experience.

This is not to say that there’s no merit to giving children some instruction on strategies that might make it easier for them to understand written texts. For instance, it’s probably useful to teach children that the conventions of fiction and non-fiction are very different. But it still remains true that children are more likely to understand either by reading lots of examples of each rather than by investing lots of time in practising generic comprehension strategies.

The sort of strategy I think it’s worth spending some time on is to explicitly teach students how to skim and scan. Skilled readers do these things without thinking but for struggling readers they’re likely to be more alien. So, for instance, it’s worth telling students that the key to unlocking the meaning of a paragraph of non-fiction text is to pick out the head noun in the topic sentence.

The Laki eruption was one of the most devastating eruptions in human history. Iceland lies on the mid-Atlantic ridge and its volcanoes pose a constant threat, although very few of them produce violent eruptions because the magma is usually basaltic and relatively free-flowing. In 1783–84, a major eruption from the Laki fissure poured out an estimated 14 km3 of basaltic lava and clouds of poisonous compounds. The volcano is located in a remote part of Iceland and no one was killed by the event itself. However, the secondary effects were devastating because the poisonous cloud killed over half of Iceland’s livestock population, leading to a famine which killed approximately a quarter of the population.

In the example above, the head noun is the phrase ‘the Laki eruption’. Skilled readers don’t even notice the definite article and are hardly slowed down by the fact they probably have no idea what Laki is. Thir attention latches on to ‘eruption’. If this were a passage of fiction the word could relate to an angry teacher but because it’s non-fiction the list of things likely to erupt is strictly limited. Even if only given two seconds to read the whole passage, skilled readers are developing a volcano hypothesis long before they actually see the word.

Struggling readers will attempt to read the passage very differently. Unless specifically told not to, they’re likely to spend as much time decoding ‘the’ as any other word and, because they’re so used to encountered unfamiliar vocabulary are likely to be completely thrown by ‘Laki’. It may be they give up before even getting to ‘eruption’ but even if they do persevere they’re less likely to be familiar with the word.

One important tip is for teachers to ensure that students know at least the head nouns before being asked to read a passage, but it’s also worth explicitly teaching how to go about skimming a text. Students who don’t read much need to be told to ignore ‘the’ and that it doesn’t matter that they don’t know what ‘Laki’ is. If they need to know it will become clear later.

There’s a similar body of knowledge we can share around scanning for information. In the extract below, try to locate the answers to the following questions:

  1. How many Irish provinces were there?
  2. When did Henry II persuaded Pope Adrian IV to give him authority to conquer Ireland?
  3. What were the names of the 3 towns taken by the English?

Six kings ruled six Irish provinces, each of which had many tribes with their own kings. Traditionally a High King of Ireland claimed tribute from the other kings. The last High King was Brian Boru, King of Munster, who died in 1014 while defeating the Danes at Clontarf. His rival, the King of Leinster, aided by the Danes who, after 1014, ruled Dublin, Waterford and Limerick. Four Kings – of Leinster, Munster, Connaught and Ulster – fought for Boru’s crown. In 1154 Henry II persuaded Pope Adrian IV to give him authority to conquer Ireland. In 1166 he had his chance. MacMurrough (Leinster) had stolen the wife of O’Rourke of Breffney, a neighbour of O’Connor (Connaught). All the kings condemned MacMurrough and banished him. He asked for Henry’s help. He sent de Clare (Strongbow), Earl of Pembroke to lead an army of chain-clad knights, supported by Welsh archers. The unarmoured Irish, with their Danish battle-axes were no match for them. The English took Wexford, Waterford and Dublin.

Easy, right? But for many struggling readers this is a daunting task. The slightly trivial sounding observation that makes it easy for skilled readers is the knowledge that the first answer is at the beginning, the second is in the middle and the third is at the end. Once you’ve located the answer to the second question there’s no way you’d go back to the beginning to look for number 3. You just know that the answer will be presented sequentially. This makes the process to scanning way for efficient for skilled readers. The reason why some children take forever to complete these kinds of comprehension activities is that they go back to the beginning of the text for every single new question!

We should tell children that the answer to a ‘how many’ question will always be a number and, in the case of a history text book, the answer to a ‘when’ question will almost always be a date rather than something vaguer like ‘last Tuesday’. And the answer to question asking for names will always result in the need to locate proper nouns. Dates and capital letters are easy to scan for because they’re like little eye-magnets.

But even more importantly we should let students know that in the context of school, the answers to comprehension questions emerge sequentially. The answer to the first question is likely to be in the first paragraph and the answer to the second question in the second paragraph. Skilled readers seem to know this instinctively, struggling readers do not.

So, is reading comprehension actually a thing, or is it all just language comprehension? I think the best answer here is to acknowledge that written text and spoken language tend to be very different and the kind of language that is used in text is likely to be very different from that used in speech. But, it remains true if you find it easier to comprehend something read aloud, then you almost have a fluency problem. Once children are fluent readers is likely to be the case that text that doesn’t make sense when read wouldn’t make sense when heard either. If this is the case the problem is due to a lack of knowledge.

To conclude, my best bets for improving children’s comprehension of written texts are as follows:

  1. Make sure children can decode at around 250 wpm. Reading fluency is a limiting factor.
  2. Spend more lesson time reading aloud with students ensuring that they encounter unfamiliar vocabulary and contexts and that these are discussed and understood.
  3. Beyond that, it’s worth giving children specific advice on how to use glossaries, headings, references and other mechanisms to make reading non-fiction easier and it’s also probably worth giving some generic insights into how to pull information out of texts.

In answer to the question is reading comprehension a thing: it probably is, but it’s definitely not the same thing many people assume it to be.

* There’s some general confusion about the difference between the speed at which we read aloud and that with which we read silently. It ought to be obvious that skilled reader will read faster when reading silently. My figure of 250 wpm is derived from a range between 238 wpm for non-fiction to 260 wpm for fiction and comes from a 2019 systematic review and meta analysis by Marc Brysbaert available as a pre print here. For comparison, Tim Roach has linked to a 2017 Reading Rockets post with data or oral reading fluency.