Mental Age Test

This test times you twice and turns the two figures into one percentage. That percentage is the speed index on the result card, and it is your speed against a reference median for people aged 18 to 24.

An index of 100 means you matched that reference. An index of 86 means you took about a sixth longer. The card then names the age band whose reference speed sits nearest your index. It reports the middle of that band as a number of years.

So the headline is not an estimate of your age. It is the answer to a much smaller question. Which of seven bands has a reference speed closest to the one you just produced?

The card carries both raw figures underneath, in milliseconds, along with how much your simple reaction wobbled from trial to trial. Those three are measurements. The band is arithmetic done to them.

There is an obvious objection to this page. It sits in the processing speed hub, and it uses neither of that hub's own paradigms. The reason is the only part of this test that was difficult.

To name an age band you need to know what a young adult scores on the same task, in a browser, on an ordinary machine. That figure barely exists. Every other test in this hub publishes no ranking at all for exactly this reason.

The digit symbol substitution test has been measured for decades, and every one of those measurements counts symbols drawn with a pencil. The symbol search test is the same story on a printed sheet. The number comparison test is a rate collected with a pen.

Reaction time is the exception. It has been measured at scale inside a browser, on tens of thousands of visitors' own machines, and reported by age band. That is why both blocks here are reaction tasks. A block built on a task with no browser anchor would add ninety seconds of work and nothing at all to the number.

Here is every step between your presses and the band, in order.

The first block gives a median simple reaction time. The reference for that is 255 ms, which is the 18 to 24 figure in this site's simple reaction model. The second block gives a median four-choice time, and its reference is 396 ms.

That second figure is the one modelling step on this page, and it is worth naming. The published four-choice model here has a median of 412 ms across all adults, and no age cohorts. The simple reaction model has both, and its young adult figure sits at 255 against an all-ages 265. Applying that same ratio to 412 gives 396.

Each block becomes a percentage of its reference. The two percentages are combined with a geometric mean, because both are ratios rather than counts. The result is read against the curve on processing speed by age. That page gives a reference median for each of seven bands, as a percentage of the youngest.

The band whose figure is nearest your index is the band you get. That is the whole method, and the weakest link in it is the 396.

1. Median simple reaction, against a reference of 255 ms.
2. Median four-choice reaction, against a reference of 396 ms.
3. Both expressed as a percentage, then combined into one index.
4. The index matched to the nearest of seven reference age bands.

The curve this page reads from publishes a middle half for every band, and those ranges run into each other badly.

The middle half of the 65 to 74 band runs from 60 to 82 per cent of a young adult median. The middle half of the 35 to 44 band runs from 78 to 106. Those two share a stretch, so a quick person in their late sixties outpaces a slow person in their late thirties.

That is not a flaw in the model. It is the most important thing the model says. Age explains a small slice of why two people differ on a timed task, and everything else about them explains the rest.

Run this test twice in one afternoon and watch what happens. Most people move a band, and some move two. A measurement that jumps twenty years between coffee and lunch is not measuring an age.

• A band two steps from your real age is ordinary and means nothing.
• The same person on a phone and on a laptop can land in different bands.
• A second run is usually quicker, because you now know the task.

The phrase has a real history and it was abandoned for good reasons. Early intelligence testing scored a child by the age at which a typical child answered the same items. Divide that by real age and you had a ratio, which is where the idea of an IQ number began.

It fell apart on adults. There is no age at which a typical adult stops being able to do the items, so the ratio stops meaning anything after about sixteen. No current clinical measure uses a mental age, and none has for decades.

What this page reports is narrower and more honest. It is a speed index, matched to a curve. It is not an intelligence score, it is not a mental age in the historical sense, and it is not a brain age.

Brain age scores sold as products are marketing. In 2016 the FTC fined one brain training company two million dollars over claims of this kind. Nothing on this page detects, screens for or rules out any condition, and no figure here is a diagnosis. A real worry about a change in yourself belongs with a doctor.

A reaction time collected in a browser is not just your nervous system. Roughly a third of it belongs to your screen, your mouse and the operating system underneath them.

A 60 Hz laptop and a 144 Hz monitor are worth tens of milliseconds between them, and tens of milliseconds is a band. A phone is worse. Block two is tapped rather than typed, and a thumb travelling to a square costs time a key press does not.

Sleep, caffeine, time of day and whether a message just arrived all move the figure more than a decade of ageing does. What counts as a good reaction time works through how much of any such number is equipment rather than person.

So the comparison worth having is you against your own earlier runs, on the same machine, at the same time of day. That takes the hardware out. It is also the only comparison on this page that a change in you would actually show up in.

The milliseconds under the band are the part of this page that is a measurement, and they are worth more than the band is.

Your simple reaction is spotting and moving, with no decision in it. Your four-choice time is the same chain with a decision added. The gap between them is what deciding between four options costs you, and it is usually well over a hundred milliseconds.

The trial-to-trial spread is the third figure and the least familiar. It is how far your ten simple trials scattered around their own middle. A large spread with a quick median usually means attention wandered on a trial or two rather than that you are slow.

If those interest you more than the band does, that is the right reaction. The simple reaction time test and the four-choice version both run longer, and both report a steadier number.