Why Smart People Are Bad at Mental Maths — The Three Causes
You can follow a complicated argument, write a sharp report, and hold your own in a meeting full of experts. Then someone asks you to work out 15 percent of 68 out loud, and your mind goes completely white.
Maybe you laugh it off. "I'm just not a numbers person." It is one of the few kinds of incompetence people admit freely. But privately it bothers you. How can a mind that handles abstract ideas so well fall apart at the bill-splitting stage of dinner?
The research has a more interesting answer than "some people just can't do maths." There are three different explanations, and they often get confused with one another: anxiety that eats working memory, arithmetic that was never made automatic, and, much less often, dyscalculia. They can look the same from the outside. Underneath, they are different problems with different fixes.
Why are smart people bad at mental maths?
Start with what mental arithmetic actually asks of your brain. Working out 47 + 38 in your head is not one act of thinking. You add the units, hold the 5, carry the 1, add the tens, then put the pieces back together. All of this happens in working memory, the small mental workspace that holds information while you use it.
Nicole DeStefano and Jo-Anne LeFevre at Carleton University reviewed the research on working memory in mental arithmetic in 2004. They concluded that every major part of Alan Baddeley's working memory model plays a role. That includes the central executive, which manages the steps, and the phonological loop, the inner voice that holds intermediate totals like "fifteen, carry one."
This explains much of the puzzle. Mental maths is not a test of how clever you are. It is a narrow test of how much you can juggle at once, how quickly you can pull stored facts from memory, and how calm you stay while doing it. A person can be excellent at reasoning and still be weak on any one of those.
The size of the problem is not small, either. In England's 2011 Skills for Life survey, 49 percent of working-age adults had numeracy at or below the level expected of a child aged 9 to 11. National Numeracy's later survey put the figure at 54 percent.
Cause one — maths anxiety and the working memory squeeze
The best-known explanation comes from Mark Ashcraft, whose lab spent years studying why anxious people perform worse at arithmetic. In a 2001 study in the Journal of Experimental Psychology: General, Ashcraft and Elizabeth Kirk found that people with high maths anxiety had smaller working memory spans, especially on number-based span tasks. Their errors spiked on problems that needed carrying, which are exactly the problems that load working memory the most.
Crucially, they described this as a temporary disruption, not a permanent limit. Worry takes up the mental workspace that the sum needs. When the worry goes, the space comes back.
Sian Beilock, a cognitive scientist then at the University of Chicago, took this further. In a 2005 Psychological Science paper titled "When High-Powered People Fail," Beilock and Thomas Carr gave students difficult maths problems under low and high pressure. Under low pressure, people with high working memory beat people with low working memory, as you would expect. Under pressure, their advantage disappeared. Only the high working memory group got worse, and only on the problems that made the heaviest demands on working memory.
In other words, the people most likely to "choke" on mental maths were the ones with the most mental horsepower. Their usual strategies depended on a large working memory, and pressure took exactly that away.
"Working memory is one of the major building blocks of IQ." — Sian Beilock, University of Chicago
The same pattern shows up early in life. In a study of first and second graders, Gerardo Ramirez, Beilock and colleagues found that maths anxiety hurt achievement mainly in children with higher working memory. A 2016 follow-up in the Journal of Experimental Child Psychology explained why. Anxious children used fewer of the advanced strategies, such as breaking numbers apart or retrieving facts directly, that high working memory normally allows. Children with lower working memory were already using simpler methods like counting on their fingers, so anxiety had less to disrupt.
The anxiety can also start before any maths happens. In a 2012 brain imaging study in PLOS ONE, Ian Lyons and Beilock found that in people with high maths anxiety, simply expecting a maths task activated the posterior insula, a region linked to the sense of threat and physical pain. The effect showed up while people were waiting for the maths, not while they were doing it. The study was small, with 28 adults, but it helps explain why many people avoid mental arithmetic altogether.
How common is maths anxiety in adults?
You will often see the claim that 17 percent of people have maths anxiety. That figure is usually misread. It comes from Ashcraft and Alex Moore's 2009 work with US college students, where 17 percent was simply the share of people scoring more than one standard deviation above average, a statistical cutoff rather than a prevalence rate.
Better adult estimates are lower. In an online sample of about 1,000 adults, Christopher Hart and Colleen Ganley found that 5.4 percent reported substantial maths anxiety. Milder unease is much more common. In the OECD's PISA 2012 survey, 59 percent of 15-year-olds said they often worry that maths classes will be difficult for them, and about one in three said they feel anxious when faced with a maths problem.
The link between anxiety and performance is real but moderate. A 2021 meta-analysis in Psychological Bulletin by Carlos Barroso and colleagues pooled 747 effect sizes and found a correlation of -.28 between maths anxiety and maths achievement. That is meaningful, but it also means anxiety is part of the story for many people, not the whole of it.
Cause two — arithmetic that was never made automatic
The second explanation is less dramatic. You may simply never have made basic arithmetic automatic, or you may have let it fade.
Skilled mental arithmetic depends on direct retrieval: seeing 7 × 8 and knowing 56 without working it out. Robert Siegler at Carnegie Mellon spent decades showing that people choose between strategies, retrieving an answer when the memory is strong and falling back on slower procedures when it is not. LeFevre and her colleagues challenged the old assumption that adults simply retrieve everything. Their work showed that adults, like children, still use counting and step-by-step procedures on many simple problems. Subtraction turned out to be the most procedural operation of all.
Every procedure costs working memory. If 8 + 7 is a fact you retrieve, it costs almost nothing. If you have to work it out through 8 + 2 + 5, it takes up space you needed for the rest of the problem. Two people with the same working memory capacity can look very different at mental maths simply because one of them has more facts stored ready to use.
This is where the "smart people" pattern often comes from. Bright students tend to move quickly to topics where calculators are allowed and the focus is on concepts. Their reasoning keeps growing while their arithmetic fluency stays at whatever level it reached at age eleven, then slowly gets worse through lack of use.
Calculators are often blamed, but the evidence does not support a simple story. Ray Hembree and Donald Dessart's 1986 meta-analysis, and a later one by Aimee Ellington, found that calculator use in school generally maintained or improved operational skills rather than damaging them. Calculators do not erase your ability. They just remove the daily practice that keeps retrieval quick.
Cause three — could it be dyscalculia?
The third possibility is dyscalculia, a specific learning difference in how the brain represents quantity. It affects an estimated 3 to 7 percent of people, which is about as common as dyslexia, and it is far less often identified. We cover the signs in detail in Do I Have Dyscalculia? What Number Blindness Feels Like.
What matters here is how it differs from the first two causes. A large 2018 study by Amy Devine, Dénes Szűcs and colleagues at Cambridge found that most children with high maths anxiety actually had typical or above-average maths performance. Anxiety and dyscalculia overlap, but they are largely separate conditions.
The practical difference is this. Maths anxiety gets worse under pressure and eases when you are alone and relaxed. A practice gap improves fairly quickly with a few weeks of targeted drilling. Dyscalculia stays the same in every setting and goes deeper than calculation. It affects your basic sense of quantity, such as telling at a glance whether there are four or five items, estimating amounts, and judging time and money. If you can do the sum perfectly at home but freeze when someone is watching, dyscalculia is unlikely to be the main explanation.
The cognitive dimensions behind mental arithmetic
At CognitionType, we look at a difficulty like this through the cognitive dimensions it draws on, rather than through a single label. For mental maths, three dimensions do most of the work.
Memory and sequencing. This is the core of the task. Mental arithmetic means holding numbers, keeping steps in order, and storing partial results while you work on the next part. Someone whose working memory is limited, or easily disrupted, will feel the strain on multi-step sums long before simple ones. That matches the carry-problem effect Ashcraft found. For more on this system, see our piece on working memory as the skill behind everything.
Phonemic processing. This is the dimension most people do not expect. Stanislas Dehaene's triple code model proposes that the brain stores number facts in several formats, and that rote facts such as multiplication tables are stored largely in verbal form, as chanted sequences of sounds. Holding "fifteen, carry one" also depends on the inner voice. This helps explain why people with phonological differences, including many people with dyslexia, often struggle specifically with times tables and spoken arithmetic while reasoning well about quantity on paper.
Emotional regulation. Beilock's work shows that how you handle stress in the moment directly affects how much working memory you have available. Someone who stays steady under pressure keeps their full capacity. Someone whose stress response rises quickly loses capacity at exactly the moment they need it. That is why the same person can be fluent at the kitchen table and blank at the restaurant.
Looking at these dimensions together explains why the "smart person who can't do mental maths" is so common. A strong reasoning profile can sit alongside a phonological loop that drops numbers, a stress response that takes over quickly, or both. None of that is a verdict on intelligence.
How to tell which one is affecting you
A few questions can help you work out which cause matters most for you.
Does the setting change your performance? If you can do a sum easily alone but not with someone watching, or not under time pressure, anxiety is probably doing most of the damage.
Is it slow but accurate? If you get the right answer through careful step-by-step work, the issue is more likely missing fact fluency than a problem with understanding numbers.
Is the difficulty broad and consistent? If estimating, reading clocks, handling change and judging quantities have always been hard, in every setting and since childhood, dyscalculia is worth investigating.
Are spoken numbers harder than written ones? If you lose numbers that are said aloud but can work them out on paper, the strain may sit more with phonological and working memory systems than with number sense itself.
More than one often applies. Anxiety frequently grows on top of a practice gap, and dyscalculia almost always brings some anxiety with it.
What you can do about being bad at mental maths
Write the worry down first. In a 2011 Science study, Ramirez and Beilock had students spend ten minutes writing about their worries before a high-pressure maths test. Control students' accuracy dropped by 12 percent under pressure. Those who wrote improved by 5 percent. Other researchers have questioned whether the approach helps people who are not anxious, so treat it as a tool for anxious moments rather than a general fix.
Reframe the pressure. Beilock and her colleagues have suggested that treating a stressful moment as a challenge rather than a threat can soften its effect on performance. Noticing a racing heart and reading it as readiness rather than danger is a small shift, but it targets the same stress response that eats working memory.
Rebuild fact fluency deliberately. If the problem is practice, short daily retrieval practice works better than occasional long sessions. Aim for automatic recall of number bonds to 20 and the times tables up to 10. Every fact you can retrieve instantly frees working memory for the rest of the problem.
Offload without shame. Writing intermediate steps on paper, or using your phone's calculator, is not cheating. It moves the work out of the system that is under strain.
Watch what you pass on. Erin Maloney and colleagues followed 529 children and their parents in a 2015 Psychological Science study. When maths-anxious parents often helped with homework, their children learned less maths over the school year and became more anxious about it. If you are anxious about maths, how you talk about it in front of children matters.
Understand your own profile. Because the same difficulty can come from working memory, phonological processing or emotional regulation, knowing which dimension is doing most of the work makes it easier to choose the right strategy. CognitionType can be a useful way to understand your own cognitive profile across these dimensions. It works alongside, not instead of, a formal assessment from an educational psychologist if you suspect dyscalculia.
The bottom line
Being bad at mental maths is not proof that you are less intelligent than you thought. It is a sign that one of the specific systems mental arithmetic depends on is overloaded. For most capable adults, that means working memory taken up by anxiety, arithmetic facts that were never made automatic, or both. For a smaller group, it reflects a real difference in how the brain represents quantity.
The irony in Beilock's research is that the strongest working memories can be the most vulnerable, because they rely on the very resource that pressure takes away. "I'm not a maths person" is a description of a moment, not a fixed trait, and a moment can change.
CognitionType is an informational assessment, not a clinical diagnosis. If you suspect dyscalculia, maths anxiety severe enough to affect your daily life, or another learning difference, we encourage you to seek formal evaluation from a qualified educational psychologist or clinician.