Working Memory Comes First

What the research found

Two skills sit near the centre of most conversations about attention. One is working memory: the ability to hold information in mind and work with it, such as keeping a set of instructions in your head while you carry them out. The other is inhibitory control, sometimes called self-control: the ability to stop a first impulse and choose a better response. Both are weaker, on average, in children with ADHD. The hard question is which one drives the other. This study set out to answer it.

The design is what makes it worth reading. Rather than measure both skills and look for correlations, Kofler and colleagues ran an experiment. They tested 110 carefully assessed children aged 8 to 13, a mixed clinical group of 59 with ADHD, 28 with ADHD and anxiety, and 23 with anxiety alone. Each child completed four computer tasks arranged so the researchers could turn the demand on each skill up or down independently. In effect, they could deliberately overload a child’s working memory and watch what happened to self-control, then reverse the arrangement and overload self-control to watch what happened to working memory. This is called a double dissociation, and it lets you test cause rather than mere association.

The first result was large and clear. When the children’s working memory was occupied, their self-control got noticeably worse. Responses slowed, with an effect size of d=1.53, and accuracy fell, with an effect size of d=0.67. An effect size of this kind describes how big a change is relative to the normal spread between children; anything above about 0.8 is generally considered large. The statistical support was overwhelming. The Bayes factor, a measure of how strongly the data favour one explanation over another, ran into the billions for the slowing effect. In plain terms, filling up working memory reliably broke down self-control.

The second result was the mirror image, and just as important. When the researchers loaded the self-control demand instead, working memory held steady. Here the evidence actually favoured no effect, with a Bayes factor of 18.52 pointing towards the null. That number means the data were roughly nineteen times more consistent with “self-control does not drive working memory” than with the opposite. So the influence ran in one direction only.

A third, smaller finding is worth noting. Under heavy working memory load, the children with ADHD lost accuracy more than the anxiety-only group did (d=0.50), which fits the idea that this bottleneck is especially relevant to ADHD rather than to childhood difficulties in general.

Taken together, the authors conclude that working memory sits upstream. Weak working memory appears to generate the self-control problems, and not the reverse. That is a specific, testable claim, and this experiment supports it cleanly.

What it means for you and your child

The practical message is about where to look first. If your child struggles to inhibit a blurted answer, to wait, or to stop an off-task impulse, this study suggests the visible problem may not be the root problem. The root may be a working memory system that is already full. A child who cannot hold the goal in mind has little left over to resist the wrong response. The impulsivity you see could be a downstream symptom of an upstream shortage.

That reframing changes how some everyday moments read. A child who forgets the second half of a two-part instruction, or who does the wrong step of a routine they clearly know, may not be careless or defiant. Their working memory may simply have run out of room mid-task. Reducing what a child has to hold in mind at once, by breaking instructions into single steps, writing them down, or checking in partway through, is a reasonable response to what this research describes. It lightens the load at the point where the load appears to matter most.

It is worth being precise about what this single study does and does not establish. It shows, under controlled experimental conditions, that working memory load causes self-control to falter within a task. That is genuine causal evidence, which is rarer and more valuable than correlation. But it was measured over minutes in a laboratory. It does not prove that training working memory will produce lasting real-world gains in self-control, nor that working memory is the only cause of every child’s difficulties. Those are separate questions, and the honest answer is that they remain open.

Two limitations from the authors deserve a place in any parent’s reading. The comparison group was children with anxiety rather than children with no diagnosis, a choice forced by funding, so the study cannot say precisely how these children compare with typically developing peers. And the study was not pre-registered, meaning the analysis plan was not locked in advance; this is a normal caveat, not a fatal flaw, but it is a reason to want replication before treating the finding as settled.

For a family, none of this points to a specific product or programme. It points to a way of thinking. When self-control seems to be the issue, it is worth asking what your child is being asked to hold in mind at the same time, and whether easing that hidden burden helps. This is one well-built study, and its direction is encouraging. It is not, on its own, the last word.