Spatial Reasoning Tests
Try five visual tasks that use pictures instead of words, with each one measuring a different spatial skill.
Spatial reasoning is the part of thinking that runs on pictures. You turn a shape, fold a sheet, or look for one outline hidden inside another. None of that passes through words on the way.
The five tests here ask for it in five different ways, and each one ends on a different kind of number.
The mental rotation test puts two figures made of blocks side by side, one of them turned, and prices what every quarter turn costs you. The paper folding test folds a sheet, punches through the whole stack, and counts the holes you put in the wrong place.
The embedded figures test buries a simple shape in a pattern of lines and times the hunt. The block counting test draws a pile of cubes with some of them out of sight. It takes a typed number and reports how far out you were.
The spatial ability test is the mixed paper. Count a stack of cubes, match the view from the front, then find the flat sheet that folds into a box.
Two of the five run a clock and three do not, and the split is deliberate. On rotation the time is the result, because the finding being reproduced is how the time grows with the angle. On a punched sheet the mistake is the result, and hurrying somebody only makes that measure worse.
One of the five places your number against a distribution, and that curve is drawn from published research rather than from our visitors. The other four rank you against nobody at all, and each says so on the page rather than borrowing a curve that nearly fits.
So take more than one of them. What your five numbers disagree about is more interesting than any of them alone.
The 5 tests in this domain
Numbers and reading in this domain
Turning, folding and finding are not one skill
These five pages look like five versions of the same test, and they come apart the moment you take them.
Rotation is a rate. The work is turning something at a steady speed until it lines up, and the number is how much each quarter turn costs.
Folding is a construction. Nothing turns, and you build a picture that has never been on the screen out of a rule applied over and over.
Finding a hidden figure is neither. The shape is already in front of you at the right size and the right angle. The job is getting your eye to accept a corner that is drawn as a crossing.
Counting a pile of cubes is closer to arithmetic than to any of the three. You read a height off every column and hold a running total. The cubes the picture never drew are the ones that decide the answer.
The mixed paper is closer to an exam than to any of them, because it also asks what you know. A stack has cubes behind the cubes. Eleven of the 35 sheets of six squares close into a box, and the other 24 fold onto themselves.
Published work has been splitting spatial ability into more than one factor for the better part of a century, and the list has never settled. Being good at one of these and ordinary at the other four is the usual result rather than a strange one.
Why four of these five pages rank nobody
Every other hub here places your score against a named distribution. This one manages it on a single page, and the reason is different each time.
The paper folding test people have norms for is a mark out of twenty on five way multiple choice. Its tables are scanned pages inside a manual, so nobody can open the source and check them. A mark on that test is not an error count on this one.
The published embedded figures times were taken with a stylus. The simple shape was shown, taken away, and then traced from memory inside the busy one. Our version leaves the shape on screen and takes one tap, which removes most of what the paper version was timing.
The mixed paper offers four answers an item, so a quarter of the marks are inside reach of a coin. A percentage only means something against the exact paper it came from.
Cube counting has been an aptitude item for a century without anybody publishing a distribution of the item on its own. The figures that exist are shares of multiple-choice marks on whole papers, and none of them converts into cubes out per pile.
Mental rotation is the exception, and it is a reference model rather than our data. It says so wherever it appears, and it carries no sample size, because a model does not have one.
What your screen and your mouse add
Two of these pages report seconds, and a browser cannot measure seconds cleanly. A press has to travel through your mouse and your operating system before the page hears about it.
That overhead is why the rotation page reports a gradient rather than a raw time. The gradient is the difference between your slow answers and your fast ones. A fixed overhead sits in both, so it cancels out of the subtraction.
The embedded figures number has no such protection. It is a raw time and it carries the press with it. That press is a fifth of a second or so on top of the search.
The three untimed pages sidestep the problem entirely. Nothing they report depends on when the browser noticed you.
The methodology page sets out how the timing is done, including why a stimulus is timestamped in the frame after the one that paints it.
What none of these numbers mean
A good score here means you did well at a small drawing task in a browser, on this screen, on this day. It describes performance on this drawing task only.
Spatial tasks respond to practice more obviously than most, and the second run is usually better than the first. What improves is the task you practised.
- Your screen size and your pointer are inside both of the timed figures.
- A run taken in a hurry and a run taken carefully are two different measurements.
- Your own earlier runs are the fairest comparison four of these pages can offer.
- Try another visual skill test or repeat this one under the same conditions.