I have said it once. Get your shoes and your water bottle. My son nods, turns, and walks toward his room. Thirty seconds later he is standing in the hallway holding one shoe, looking at me like I spoke a sentence he is never heard before. Where is the water bottle. What water bottle.
For a long time I read this as not listening, or not trying, or being somewhere else in his head. I don’t think that anymore.
Working memory in children this age has a real, physical limit, and the second half of my instruction fell out of it before his feet finished carrying him toward the first half.
Educational content only. This article is built from CDC, AAP, and peer-reviewed research. It is not medical advice and does not replace your pediatrician’s guidance. If you have concerns about your child’s development, your doctor is the right first conversation, not this article.
Here’s the distinction that changed how I see this whole thing. When we say a child has a good memory or a bad memory, we usually mean long-term memory: the storage system that holds birthdays, names, the words to a favorite song, the layout of a room visited once. Working memory does something else entirely. It’s the small, active workspace where information gets held only long enough to be used, not stored.
The distinction between these two systems goes back to a foundational 1974 framework proposed by psychologists Alan Baddeley and Graham Hitch, who separated the temporary, active holding of information from its longer-term storage. It still holds up decades later, and it explains something I used to find confusing: my son can recite every dinosaur name he owns a toy of, in order, without missing one, and still lose the second half of a two-part instruction ninety seconds after hearing it. Both things are true about the same child at the same time, because these are two different systems doing two different jobs.
I think of it less like a filing cabinet and more like a kitchen counter. Long-term memory is the cabinet, where things get put away and stay put, sometimes for years. Working memory is the counter, the space where you set down what you’re actively using right now, before you either file it away or hand it off to the next step. A young child’s counter is small. Things slide off the edge while he’s still holding the first item in his hands.
This is the part I misread the longest, the part where I assumed he was choosing not to listen. Researchers Hrag Pailian, Melissa Libertus, Lisa Feigenson, and Justin Halberda tested seventy-two children between the ages of three and eight, along with a group of adults, using a task that measured how much visual information each person could actively hold and search through at once. The study, published in the journal Attention, Perception, and Psychophysics in 2016, was built specifically to separate working memory capacity from everything else that tends to improve alongside it during these years: attention span, visual processing speed, and general decision-making. Once they accounted for those other factors separately, they still found a clear, steady increase in working memory capacity across childhood, one that didn’t level off until somewhere between ages six and eight, when it started to resemble what adults could do.
That separation matters more than it might seem. It means the workspace itself grows with age, apart from all those other skills getting sharper too. My three-year-old wasn’t an inattentive adult with a shorter fuse. He had a smaller counter, full stop, regardless of how hard he was trying or how much he cared about getting it right.
That rules out the story I used to tell myself in the hallway with one shoe in my hand: that he would have remembered if he’d cared more, or focused harder, or wasn’t distracted by something shinier down the corridor. The capacity limit sits underneath attention and motivation, not on top of them. You can have his full, genuine attention and still watch the second half of a two-part instruction disappear before he reaches the door.
Earlier research by developmental psychologist Vanessa Simmering, testing children at ages three, four, five, and seven, found that working memory capacity sat at around two to three items for children between three and five years old, and grew to roughly four items by around age seven. The Pailian team’s later work extended that picture forward, showing the climb continuing past seven, not leveling off until somewhere between six and eight years old, consistent with a gradual, protracted developmental climb, not an early plateau.
Put together, this told me my five-year-old forgetting the water bottle wasn’t behind schedule. He was close to exactly where a five-year-old’s working memory capacity is expected to be: a good year or two out from the point where his workspace starts to resemble an adult’s, and further still from the point where it fully does.
Knowing that number changed what a lost second step meant to me. It stopped being a red flag. It became the current size of the counter, doing precisely what a counter that size does when two things are set down on it at once.
Psychologist Vanessa Simmering, testing children at ages three, four, five, and seven, found that working memory capacity sat at around two to three items for children between three and five years old, and grew to roughly four items by around age seven.
Once the size limit felt real to me instead of theoretical, the practical adjustment was direct. Instead of loading two items onto the counter at once and hoping both survive the walk to the bedroom, I hand them over one at a time, in sequence, with a brief pause for the first one to happen before the second one arrives.
Here’s what vague used to look like in my house: Get your shoes and your water bottle, said once, at the door, while I was already reaching for my own bag. Here’s what specific and sequenced looks like now: Go get your shoes, said alone, waited out, followed by Now get your water bottle once the shoes are on his feet. Give one instruction at a time instead of stacking two or more items into a single request, and wait for the first to be completed before giving the second. The second version asks the workspace to hold one thing at a time instead of two, and the counter is no longer being asked to carry more weight than it was built for at this age.
This is about matching the size of the request to the size of the workspace doing the holding, not about lowering what I expect from him. The same child who drops the second half of a two-part instruction can usually complete a one-part instruction cleanly, nearly every time, because that instruction was never competing with anything else for space on the counter. Pointing while I speak, or walking partway with him for the first request, gives his workspace an anchor to hold onto instead of only words, which tend to slide off a small counter faster than something he can also see.
For most children, a small working memory workspace at this age isn’t a concern. It’s the expected shape of a system still under construction, and it grows on its own over the next several years without anything needing to be fixed. The Four Years Old milestone piece and the First Five Years of Cognitive Development article both touch on how this particular skill fits alongside everything else developing at the same time, if a fuller picture of the whole stage is useful to you.
What’s worth bringing up with your pediatrician is a different pattern entirely: if even single, one-part instructions consistently don’t land, if the gap between what your child can hold onto seems unusually far behind other children his age across multiple settings, not only at home, or if it comes bundled with other concerns around attention or language. Raising a Developmental Concern With Your Pediatrician walks through how to bring an observation like that into a well visit in a way that gets taken seriously. Your pediatrician is the right person to weigh any pattern like this against your child’s overall development, and there’s no harm in simply asking the question at your next visit if something about it doesn’t sit right with you.
Most of the time, though, the shoe in the hallway isn’t a warning sign. It’s a small counter doing exactly what small counters do, on its way to getting bigger.
Working memory in children this age has a limited capacity for holding information actively in mind, separate from long-term memory. A 2016 study by Pailian and colleagues, published in Attention, Perception, and Psychophysics, tested children ages three to eight and found that this capacity increases steadily with age and does not approach adult-like levels until somewhere between six and eight years old. A child who forgets the second part of a two-step instruction is typically not being inattentive or defiant. The workspace holding the instruction is simply smaller at this age than it will be later.
No. Working memory is the active workspace used to hold and manipulate information in the moment, while long-term memory is the separate storage system for facts, routines, and experiences accumulated over time. A child can have a strong long-term memory for names, songs, or routines while still having a limited working memory capacity, because these are different systems governed by different developmental timelines.
Research by Pailian, Libertus, Feigenson, and Halberda (2016) found that working memory capacity approaches adult-like levels somewhere between ages six and eight, after controlling for improvements in attention, perception, and decision-making that also occur during these years. Earlier research by Simmering (2012) found capacity around two to three items at ages three to five, growing to roughly four items by around age seven, consistent with a gradual, protracted developmental climb, not an early plateau.
Give one instruction at a time instead of stacking two or more items into a single request. Wait for the first instruction to be completed before giving the second. Pointing to the object involved, or walking partway with the child for the first step, gives the working memory workspace a visual anchor to hold onto in addition to the spoken words, which reduces the chance that the instruction is lost partway through.
A small working memory capacity at this age is expected and not, by itself, a reason for concern. It is worth raising with your pediatrician if even single, one-part instructions consistently do not land, if the gap between what your child can hold onto seems noticeably wider than same-age peers across multiple settings, or if it appears alongside other concerns involving attention or language. Your pediatrician can weigh the full pattern against your child’s overall development.
Share this article
