Working Memory vs. Short-Term Memory — What's the Difference?

The gap between how many digits you can repeat and how many you can reverse shows what happens when storage has to work under extra control demands.

Feel the Difference Below (Forward vs. Backward Digit Span) ↓

Read these digits once, look away, and say them aloud: 7 — 2 — 9 — 4.

Easy. Now four fresh ones — 8 — 3 — 6 — 1 — but say them backward.

Same quantity of information. Same handful of seconds. Yet the second version is heavier, slower, and gives out sooner — even though the amount you had to keep hold of never changed. Something other than the quantity is doing that.

One Holds. The Other Holds and Works.

Short-term memory is the holding function: keep a small amount of material available, unchanged, for a matter of seconds. Repeating a phone number, carrying a door code across a hallway, echoing back the end of a sentence. Nothing happens to the material — it goes in and comes out the same.

Working memory is that holding plus a workshop around it: comparing, reordering, updating, combining, all while the material is still being held. Mental arithmetic needs it. So does following directions with more than two steps, or untangling a sentence with several clauses stacked inside it — which is why most of Cognitive Train's free brain training exercises load this side rather than plain storage.

The reversal task is a useful illustration because the quantity stays fixed while the demands change. You still hold four digits — and you now have to reorder them without losing them, which draws on control and sequencing that plain repetition never asks for.

Why Psychology Needed a Second Word

Short-term memory came first, as the middle stage of the 1960s memory pipeline — one buffer sitting between the senses and permanent storage, doing all the temporary work.

Trouble arrived when researchers asked whether that buffer really was the workspace of thought. In Alan Baddeley and Graham Hitch's 1974 experiments, people held digit loads in mind while simultaneously reasoning or comprehending sentences. If a single store handled both, filling it with digits should have wrecked the reasoning.

It didn't. Performance slowed, but it held up far better than a one-box model predicted.

The conclusion reshaped the field: temporary memory isn't a passive container but a multi-part system, with separate storage for verbal and visual material plus an attentional executive coordinating them. Storage could be loaded while the executive carried on reasoning — which is precisely what the dual-task results showed. "Short-term memory" stayed as the name for the storage function. "Working memory" became the name for the whole apparatus.

What the Gap Does and Doesn't Tell You

Now back to the digits — carefully, because this is where popular accounts overreach.

Forward span emphasises brief storage. Backward span uses that same storage while adding reordering and control demands on top. The drop between them reflects those added demands, and that's a genuinely useful thing to feel.

What the drop is not is a clean subtraction leaving a pure measure of one executive component. Backward recall still leans heavily on storage, and researchers continue to disagree about exactly which processes it taps — people use different strategies to reconstruct an order, and the sequencing itself makes demands of its own. So treat the two scores as "storage" and "storage under extra control demands," not as shelf and manager cleanly separated.

The size of the gap varies by person, by how the task is administered, and by population, so it isn't a number to expect yourself to reproduce. The direction is consistent enough to notice, which is the part worth having. When you run the digit span test both ways, you aren't measuring one ability twice.

Forward span leans on storage. Backward span adds reordering and control. Compare both of yours below ↓

Where the Line Gets Blurry

One admission is owed here, because it's the real reason these terms get tangled: researchers themselves draw the boundary differently.

Nelson Cowan's 2008 analysis of the terminology maps the landscape. Some researchers use the two terms nearly interchangeably. Some reserve "working memory" strictly for tasks combining storage with processing. In Cowan's own framework, working memory is short-term storage plus the attention that manages it — which makes short-term memory a component nested inside working memory rather than a competing system.

The nested view is the one worth carrying around. Every act of working memory contains short-term storage; not every act of short-term storage involves the workshop. Holding the number is inside reversing the number. The reverse isn't true — and the types of memory guide shows where that nested pair sits among the systems around it, with the wider brain tests collection covering the abilities that lean on both.

Why the Distinction Earns Its Keep

It matters most in what each one predicts.

Simple storage span turns out to say surprisingly little about complex thinking. Measures that add manipulation predict reading comprehension, mental arithmetic and reasoning considerably better — unsurprising, once you notice that all three are manipulation under load. The two abilities can also come apart in the same person: a solid storage span alongside real difficulty the moment reordering or updating enters the task, a profile that shows up across a range of attention and learning difficulties.

Which is why the everyday phrase "bad short-term memory" often refers to a broader working-memory or attention problem. Losing the thread mid-task, forgetting instructions while carrying out step one, arriving somewhere with the reason left behind — those are workshop failures, not shelf failures. Naming the right one points at workarounds that actually work: write intermediate results down, chunk the material, and cut the competing load rather than trying to concentrate harder. The Memory & Recall section has tests for both sides if you want to see which one is actually limiting you.

Feel the Difference Yourself

The digit span test below runs both directions back to back. Forward first — repeat the sequence as given. Then backward, where the same sequence has to be reversed before it leaves your mouth.

Watch for the moment it stops feeling like remembering and starts feeling like work. That moment is the boundary this entire article is about.

For a broader profile across several memory skills, try the Short Term Memory Test — or push the manipulation side considerably harder with the N-Back Test.