The claim and the question behind it
Headlines often say that people are reading less, or that reading is making a comeback. Both can be reported from real data. The reason is that “reading” is not one clear thing, and the figures depend on a chain of decisions: what counts, who is asked, how they are asked, and what they remember.
Understanding that chain is more useful than trying to decide which headline is right.
What counts as reading
A survey has to choose a definition, and the choice often matters more than any trend in behaviour.
- Books only, or anything? Some surveys ask about books. Others include newspapers, magazines, websites, messages and social media posts. Under the second approach, almost everyone reads daily; under the first, a smaller group does.
- Print or all formats? If e-books and audiobooks are left out, a shift in format can look like a decline in reading. If they are included, listening may or may not be counted as reading, which is itself debated.
- For pleasure or for any purpose? Study and work reading are sometimes excluded. A student reading hundreds of pages for a course may be a non-reader in a pleasure-reading survey.
- Finished or started? Counting books completed is different from counting books begun, and many readers abandon books partway through.
- Comics and fiction categories. Whether graphic novels, manga, or fan fiction count can affect results, especially among younger readers.
Because of these choices, two accurate surveys can disagree. A fall in “books read for pleasure” is consistent with an overall rise in time spent reading text on screens.
How self-reports work, and where they bend
Most data on reading comes from asking people. This is convenient and often the only option, yet it has well-known weaknesses. Self-reported behaviour and observed behaviour can differ, and reading is a case where several distortions can be at work.
Recall. People may struggle to remember how many books they read last year. Some estimate from their general sense of themselves as readers; others recall the most memorable or most recent books. Longer recall periods tend to produce rougher answers.
Social desirability. Reading is widely regarded as a good habit, so some people may overstate how much they read, or give answers that fit how they wish to be seen. Others may understate it because they do not count what they read as real reading. The direction and size of such effects vary between people and settings.
Rounding and thresholds. Answers like “about one a month” or “a few a year” are common. Responses can cluster at round numbers or at the boundaries of answer categories offered.
Interpretation of the question. Asking “Did you read a book in the last twelve months?” invites people to think about whether a half-read book, a cookbook or a textbook counts. Different respondents may decide differently, and the same person may decide differently on another day. Small changes in question wording can move the results, as discussed under question wording.
Averages hide skew. Reading is often concentrated among a minority of heavy readers. The average number of books can rise or fall with a small number of people, while the share who read none might barely move. A single average can mask both stories.
Diaries, logs and other sources
Self-reports are not the only evidence, and alternatives each carry their own limits.
- Time diaries record reading on a particular day. They avoid some recall problems but treat reading as a short main activity and may miss reading done alongside other tasks. Because a given day contains only a fraction of a person’s reading, the numbers for reading are often small.
- Library loans and sales record borrowing and purchasing. These are real transactions but reflect only one route to reading, and exclude gifts, second-hand books, sharing and free material. Titles bought are not necessarily read.
- Device and app logs from e-readers and audiobook services record pages or minutes, but only for users of that service.
- Reading tests and literacy assessments measure ability, not habit.
Combining these can offer a fuller view, but they rarely line up exactly, and differences in coverage are often as large as the differences in behaviour being studied.
Why comparisons over time are tricky
Trend claims about reading often compare surveys carried out years apart. Several things can change in that time without any real change in reading:
- The questionnaire wording or the list of formats may have been updated.
- The mode of surveying may have shifted, for example from interviews to online forms, which can affect how people answer.
- The population surveyed or the mix of respondents may differ.
- The meaning of reading in daily life may have changed as more text arrives on screens.
Comparing groups also needs care. Younger and older people may read different things in different formats, and a survey that fits one group’s habits poorly can make them look like non-readers.
Reading the claim wisely
When a story says people read more or less, it helps to check:
- What exactly was counted: books, any text, or a format?
- Was the question about the last week, month or year?
- Was the same question used in earlier surveys, with the same wording?
- Is the number an average or the share reading at all?
- Did the data come from asking people or from records?
A balanced reading of most reading statistics is that they reveal how people describe their own reading under a given definition, which is not quite the same as how much reading takes place. That is still useful information, as long as it is treated as such and not as a direct count of pages turned.