What the evidence shows, and what it does not
A review commissioned by the Nuffield Foundation pooled the trials of parent-led shared reading and found gains that were real and moderate rather than transformative. The clearest and most consistent effect was on receptive vocabulary — the words a child understands — at a standardised effect of around 0.68, which the reviewers translate to roughly eight months' advantage on the usual educational benchmarks. Pre-reading skills came out at about 0.39.
Two findings are less convenient and belong in the same breath. Expressive vocabulary — the words a child actually says — showed an effect of about 0.38 with a confidence interval that just crossed zero, which makes it a trend rather than a demonstrated result. And the measures based on parent report produced larger effects than the directly assessed ones, which the reviewers took as a sign of inflation rather than of hidden benefit: a parent who has spent six weeks reading to a child reports a more talkative child.
Three caveats matter for anyone reading to a one-year-old. The receptive gains were smaller for children under three than for children over three. None of the included studies was carried out in the UK. And every study measured outcomes immediately afterwards, so the common claim that early reading produces lasting school readiness or later attainment is not something these trials can support either way.
The honest summary is that shared reading reliably widens the set of words a small child understands, probably helps with the building blocks of reading, and has not been shown to do the larger things often claimed for it. By the standards of this subject that is a strong result.
The million-word claim
The statistic in circulation is that a child read to daily hears over a million more words before starting school than a child who is never read to. The figure is widely quoted as a finding. It is an arithmetic estimate.
The researchers counted the words in sixty commonly read picture books, assumed various reading frequencies, and multiplied across the five years before school. The headline figure depends on the top tier of that model — five books a day, every day, for five years. No child's word exposure was measured.
That does not make it meaningless. Books do contain vocabulary that rarely comes up in household conversation, which is the real mechanism and fits the receptive vocabulary finding well. But it is a projection from book word-counts, not an observation, and it should be read as an illustration rather than a measurement.
The talking matters at least as much as the text
The approach usually called dialogic reading treats the book as a prompt for conversation rather than a script: asking what is happening, waiting for an answer, following whatever the child is interested in, naming things in the pictures, and letting the child lead the direction. A page can be abandoned halfway if something on it has caught their attention.
It would be convenient to say this is proven superior to simply reading the words. The UK review could not test that, because the trials it pooled used approaches that overlapped too much to separate. What can be said is that the mechanism running through the evidence is vocabulary exposure, and a conversation about a book contains a good deal more of it than a recital does. Reading the words plainly to a child who wants to hear the story is also fine, and better than not reading.
Different voices for different characters, letting a toddler turn the pages, asking where the cat has gone — these are not refinements for people who are good at it. They are the active part.
The same book, again
Small children ask for the same book repeatedly, and adults find this trying. It is not a failure of imagination on the child's part and it is not something to steer them away from.
Repetition is how a word moves from being recognised to being known, and a familiar book is the one where a child can start to anticipate, join in, notice detail they missed, and eventually supply the next line. A child who has heard a book forty times is getting more from the fortieth reading than from a new book, even though the adult is getting considerably less. Introducing new books alongside the favourite works better than retiring it.
The toddler who will not sit still
Very young children often have no interest in sitting through a whole story, and the sensible response is to stop rather than to persist. Reading as long as the child is engaged and leaving the rest is a complete success; a book finished over a protesting toddler teaches the wrong association.
Wandering about, playing with something else, or facing the other way does not necessarily mean a child is not listening — many are, and will object if a page is skipped. A toddler who is quiet and busy is usually taking it in. What is worth avoiding is turning the whole thing into a sit-still exercise, since the only part that matters is that books are something pleasant that happens with an adult.
Books in the home language
Where English is not the main language spoken at home, reading in the home language is the better choice, not a compromise. A parent reading fluently and expressively in their stronger language gives a child far more than the same parent reading haltingly in a weaker one, and the vocabulary and narrative understanding built in one language transfer to the other rather than competing with it.
BookTrust supplies books in additional languages for families in England and Wales, and its guidance for parents is published in twenty languages, so this is provided for rather than merely permitted.
Print and screens
A printed book is the better default, and the reason is specific rather than moral. Electronic books loaded with sound effects, animations and games tend to reduce the conversation around the book, which is the part doing the work — attention goes to the device and the talk drops away. A plain digital text read together is much closer to a printed book than an interactive one is.
An audiobook is a different thing again, and a good thing, but it is not shared reading: nobody is asking the child a question or following their interest.
Free books, and the library
This is the practical part, and it is widely unclaimed.
- Bookstart Baby is free and universal for families with babies up to twelve months in England, Wales and Northern Ireland. The pack contains a book, a finger puppet and guidance on sharing stories, and is handed out through health visitors, libraries, registrars and other early years services. It is intended to arrive in the first six months.
- Bookbug bags are the equivalent in Scotland, through the Scottish Book Trust.
- Library membership is free from birth, children's borrowing carries no fines in most services, and most libraries run free story and rhyme sessions for babies and toddlers. For a child who wants the same book forty times and then never again, borrowing is a better arrangement than buying.
One honest note: the trials reviewed above tested reading interventions, not book gifting. Nobody has put a universal scheme like Bookstart through the same test, so its value rests on the reasoning rather than on trial evidence. Free books and a free library are worth having regardless.
Where reading is not the answer
Shared reading supports language development in children developing typically, and the evidence also looks positive for children with language difficulties. It is not a treatment for a language delay.
Where a child's speech or understanding seems behind, or has gone backwards, that is a conversation with a health visitor or GP rather than something to read away. A referral to speech and language therapy does not replace reading together; the two run alongside each other.
Not a performance
The last thing worth saying is that none of this needs to be done well to count. A tired adult reading three pages badly before bed is doing the thing that the evidence is about. The anxious version of this advice, in which every session is an opportunity being seized or missed, is not supported by anything above and tends to make the whole business worse.