Developmental psychologist Dr. Rachel Barr has spent her career asking whether babies and toddlers can actually learn from video, and her answer reshapes how parents should think about "educational" apps before age three.
Dr. Rachel Barr is Professor of Psychology at Georgetown University and director of the Georgetown Early Learning Project, the research group whose experiments sit behind much of what we actually know about infants and screens. While the public argument about toddlers and tablets tends to run on instinct and alarm, Barr's lab runs on something rarer: carefully designed studies of what babies and toddlers genuinely take in from a screen, and what they can do with it afterwards.
That second part is her signature. Plenty of research asks whether young children pay attention to video; they visibly do, which is part of what worries parents. Barr asks the harder and more useful question. After the watching, can the child use what they saw, with real objects, in a real room, by real hands? It is the difference between a child who has seen something and a child who has learned it, and for the youngest children that gap turns out to be wide.
The central finding from Barr's body of work is known as the transfer deficit. Show a baby or toddler a simple task on video, then hand them the real objects, and they perform far worse than children who watched the very same demonstration done live, by a person in the room. The information was on the screen. The child was watching. And yet, when the moment came to use it in the three-dimensional world, much of it did not come across. Learning from a screen, for this age group, is not merely a weaker version of learning from life. It is learning that struggles to leave the rectangle.
The source
Georgetown Early Learning Project, Georgetown University · Read the original
Why should this be? Think about what a screen strips away. A two-dimensional image carries no weight, no texture, no depth a hand can confirm; the cup on the screen is the wrong size and lives behind glass. For an adult brain, fluent in symbols, none of this matters; we translate effortlessly between a picture of a thing and the thing itself. A toddler is still building that translation. Asking a fourteen-month-old to learn a real-world skill from video is asking them to solve two problems at once: the task itself, and the mapping between the flat world and the solid one. The second problem quietly swallows much of the first.
The honest, hopeful part of the research, documented across the Early Learning Project's studies of infant learning from media, is that the deficit is not a wall, it is a slope. It shrinks with age, as children become more practised symbol-readers. It shrinks with repetition: see the same demonstration on video several times and more of it gets through. And it shrinks with co-viewing, when a parent sits alongside, points, names, and connects what is on the screen to the world the child knows. But every one of those supports is doing the same job: compensating for the gap between 2D and 3D. The live demonstration, the real person showing the real thing, needs no compensation at all. Before roughly age three, it remains the gold standard.
The first implication is a reframe of the word "educational". An app or programme aimed at a toddler can be full of genuine teaching content and still teach very little, not because the content is wrong but because the channel is. The transfer deficit means the question to ask of any screen product for an under-3 is not "is this good material?" but "can my child get this material out of the screen and into their hands?" For the youngest children, mostly they cannot, or only partially, and only with effortful help.
The second implication is gentler than it sounds. This is not evidence that screens damage babies; the deficit describes learning that fails to arrive, not harm that does. The cost is opportunity: minutes spent learning inefficiently from a flat world that could have been spent learning fluently from the solid one.
Third, the research tells you exactly what to do if a screen is going to be part of life anyway, as it is in nearly every home. Repetition helps, so a familiar programme repeated beats a stream of novelty. Co-viewing helps most of all: sit with them when you can, name what you see, and afterwards find the real version of it, the real dog, the real ball, the real cup. You are lending your child your own fluency in symbols until they grow their own. And fourth, hold the comparison the right way round. The finding is not that video teaches nothing; it is that a live person with a real object teaches far more, faster, with none of the translation tax.
Put a real wooden shape sorter in front of a sixteen-month-old and you can watch the difference the third dimension makes. The cylinder is heavier than it looks. It rolls when it is put down, which the square block does not, and the child clocks this with visible interest. The triangle goes in the round hole, fails, gets turned, fails again at a new angle, and somewhere in those failures the child's fingers are learning what no animation can show: that a shape is not a picture but a solid fact, the same on every side you cannot currently see. When the block finally drops through with a knock, the lesson has not been displayed to the child. It has happened to them, in the wrist and the fingertips, where it will still be tomorrow.
An app version of the same game presents a flat triangle and a flat hole, and a finger that drags one to the other. Whatever is being learned there, it is not the thing the real toy teaches, because there is nothing to grip, no weight to manage, no wrong angle to feel resisting the hand. Barr's transfer deficit names exactly this gap. The screen version is a picture of the lesson; the wooden version is the lesson. This is why real, hold-it-in-your-hands educational toys → are not the nostalgic option but the evidence-aligned one for the under-3s: stacking, sorting, posting and threading all live in precisely the years when the screen teaches least and the hand teaches most.
At Jaques London we have been making toys on this principle since 1795, longer than any other games company in the world. Nothing in our educational wooden toy collection → needs a screen to demonstrate it, because the demonstration is the toy: a parent posts one shape, slowly, while the child watches, and then the child's turn arrives. That little scene, an adult's hands showing a real object to a watching child, is the exact condition Barr's research identifies as the one infants learn from best. It needs no subscription and no update. It needed none in 1795 either.
The Georgetown Early Learning Project publishes its research and family-facing summaries, and is the right starting point for parents who want the evidence first-hand. For the broader picture, our guide to the science of play and screens places the transfer deficit alongside the rest of the research, our screen time guide by age turns it into practical limits, and our guide to reducing toddler screen time is the gentle, workable version of what Barr's findings imply.
Not nothing, but markedly less than from the same demonstration given live, and what is learned transfers poorly to real objects. The deficit shrinks with repetition and especially with a parent co-viewing and connecting the screen to the real world. The practical reading is not "video is poison" but "video is a low-bandwidth channel for this age group, and a person with a real object is a high-bandwidth one".
Con is too strong; mismatch is fairer. Many are made in good faith with sound content, but the transfer deficit sits between the content and the child: before roughly age three, the same teaching delivered by a person with real objects gets through far more reliably. If you use an app, use it with your child, and follow it with the real-world version of whatever it showed.
Video chat is the case most experts treat differently, because it is live, responsive and social: the person on the screen answers the child, names what the child is holding, and reacts in real time. That back-and-forth supplies some of what pre-recorded video lacks. It works best as a shared activity, with a parent alongside helping the youngest children connect the face on the screen to the real grandparent they know.
There is no single switch-on date, but the deficit eases gradually across the toddler years, with children past roughly age three increasingly able to take a lesson off a screen and use it, especially when an adult is involved. Even then, the principle holds in softer form: anything a young child can learn by doing with their hands, they will usually learn better that way.
Britain's oldest toy maker. Every toy open-ended, wooden, and built for exactly the kind of play the research recommends.