Two headlines can be true and still leave us with opposite impressions. Among children aged 8 to 18 in the UK, enjoyment of reading reached a 20-year low in 2025. In 2026, it rose. The improvement matters. So does the fact that enjoyment remains far below where it stood a decade ago.
As an editor, I know why the shorter version is tempting. It fits neatly into a headline. The fuller story takes another sentence and a reader willing to stay with it. That small act of holding two facts together is part of deep reading.
Reading closely asks us to notice the date, the scale of a change, and the qualification that keeps a true statement from becoming a misleading one. When we lose that habit, we become quicker to form an opinion than to understand a claim. A society can be surrounded by words and still struggle to read what they mean.
Deep reading is more than finding the answer
Finding the right line is easy to mistake for understanding the whole argument. The harder task is following a claim through its evidence and limits before deciding what we think.
What the PISA results can tell us
The latest international reading data give us a reason to pay attention. Across 35 OECD countries with comparable results, the average reading score of 15-year-olds fell by 25 points between 2018 and 2025. The share classified as “hasty” readers on a reading-fluency task rose from 6.6% to 11.4%. These students answered unusually quickly and made mistakes.
I find those figures troubling. They also deserve careful reading of their own. PISA measures students in participating education systems, not every adult’s reading habits. Its results cannot tell us whether screens, AI summaries, schooling, or some combination explains the decline.
The small words that change a claim
The pattern is still hard to dismiss. The assessment found persistent difficulty with longer texts. A reader can locate a fact and yet struggle to follow how several facts fit together. That second task is often where judgment begins.
Consider how often a public claim depends on a small word: may, some, among, after. A study may find an association, while a headline implies a cause. A new rule may apply to some people, while a post tells everyone to worry. Reading the sentence after the striking claim can change the entire story.
When skimming becomes our only setting
Skimming saves time, and sometimes it is exactly the right choice. Trouble begins when we use the same speed for a passing message and a claim that could change what we believe.
Skimming has a purpose
We skim because we have to. No one can give every email, article, search result, and document the same attention. Research on skimming under time pressure suggests that it can help readers retain a text’s main ideas. Details and inferences demand more attention.
That trade-off matters when a detail carries the meaning. A menu can be scanned. A medical instruction, a contract, or the evidence behind a charged public claim deserves another pass. The skill is recognizing the difference.
What fragments leave out
Our information habits make that choice harder. A headline arrives without its report. A chart arrives without its definition. A sentence from a speech circulates without the sentence before it. Each fragment may be accurate on its own and misleading in isolation.
The device is only part of the story. A book on a phone can be read with care, and a printed page can be skimmed without thought. I worry about the pace we carry from one task to the next. A habit built for clearing messages can follow us into the texts that deserve our full attention.
The temptation is to treat recognition as understanding. We have seen the subject, read a summary, and heard a confident explanation. It feels familiar. Then someone asks what the original argument actually established, and we discover how much we filled in ourselves.
When an AI summary becomes the source
AI summaries can make a dense document less intimidating. They can explain unfamiliar terms and help a reader decide where to begin. I can see why people use them.
The question is what happens when the summary becomes our only encounter with the source. In a randomized study involving 405 secondary-school students in England, those who took notes, either alone or while using an AI tool, did better on comprehension and retention tests three days later than students who used the AI tool alone.
This was a learning task with teenagers. It cannot tell us how every adult reads the news. It does show the value of actively working through material, even when a tool can explain it for us.
A useful summary gives us a route into an argument. It cannot decide which qualification will matter to our particular question, or take responsibility for a conclusion we repeat in public. When the stakes are high, I want the original text within reach.
Whoever condenses a document chooses what survives. Those choices may be sensible, careless, or self-serving. If I never return to the source, I surrender my chance to question them.
Sometimes we need to leave the page
There is another limit to reading closely. A beautifully written page can be wrong. We may follow every paragraph and still miss who paid for the message or what other evidence contradicts it.
Researchers teaching students to assess online information have found value in lateral reading: leaving a site to check its source and claims elsewhere. In one school-based experiment, students taught these strategies became better at judging questionable websites.
So the habit I want to defend has two parts. Read a consequential claim carefully enough to know what it says. Then look beyond it to learn whether it deserves your trust. Sometimes the smartest reading decision is to leave a dubious page quickly and spend your attention on a better source.
Without those habits, public disagreement can start before anyone has established the facts of the disagreement. One person responds to a headline. Another defends a paragraph neither has read. A correction arrives with context, but the original impression is already simpler and easier to pass along.
The public cost of missing the fine print
A mistaken reading can become a shared assumption before anyone returns to the original. That matters most when the words shape public decisions or how we judge one another.
Who explains public decisions?
Public decisions often live in documents few people have time to study. A school rule, a budget proposal, or a government notice can turn on who is included, when a change takes effect, and which exceptions apply. We need journalists and other people who can explain those documents. We also need enough confidence in our own reading to question an explanation when it seems too neat.
Otherwise, the person who selects a few lines for us gains influence over the whole argument. We may still have access to the original document, but access alone gives us little if we never examine what was left out. The same is true when an AI tool produces the summary.
What fair disagreement requires
Close reading matters in disagreement, too. If I am going to challenge a position, I owe its author an accurate account of it. A clipped sentence makes an easy target. The full argument may reveal a qualification I dislike, a concern I share, or an assumption that deserves a better challenge. That patience will not end disagreement. It can make disagreement more honest.
I regard these as civic risks. A reading test cannot measure their size or establish that they cause distrust or polarization. But a shared argument becomes difficult when we cannot first agree on what was written.
A reading culture is built, not merely requested
Calls to “read more” can sound easy from a desk with books, time, and a reliable connection. Long workdays, difficult public language, and constant demands on attention also shape what people can read carefully. In Bangladesh, where this question has an immediate educational dimension, UNICEF’s 2025 survey found that only 50.2% of children aged 7 to 14 demonstrated foundational reading skills. That is a measure of basic skills, not a trend in deep reading. It is a reminder that the opportunity to read closely must be built from the beginning.
Children need effective instruction and access to material they want to read. Adults need public information written clearly enough to examine. Schools can teach students to follow an argument and check a source. Those are practical skills, not a demand that every child become a literary scholar.
Journalism has obligations here, too. We cannot ask readers to notice a caveat that our headline erases. We cannot complain that people misunderstand evidence while hiding its limits far below the opening. Clear headlines, accessible context, and links to underlying material give readers a fair chance to judge for themselves.
I do not expect anyone to read every word they encounter. I hope we can preserve the judgment to recognize which words matter. Read past the arresting line. Check the date and the denominator. Ask whether “linked to” has quietly become “caused by.” Find out who is speaking.
A society that stops reading closely will not fall silent. It will keep arguing, often over versions of a text that someone else selected for it. I would rather begin with the words themselves, even when they make the argument harder.
