Paper Folding Test

Fold a sheet of paper in half, punch a hole through it, then open it out. There are two holes, not one.

The second one sits at the mirror of the first, and the crease is the mirror line. That is the whole rule of this task, applied over and over.

Fold twice and the punch passes through four layers, so four holes appear. Fold three times and it passes through eight.

Every fold on this page is a fold in half, so every fold doubles the layers and doubles the holes. The strip of pictures shows each one as it happens.

The sheets are drawn at one scale, which is why they shrink across the strip. The size of the last picture tells you how much paper the punch went through.

A sheet folded three times really is an eighth of the one it started as. That is not a drawing trick, it is what folding does.

The headline is the number of holes you put in the wrong place, averaged over the ten sheets. Lower is better and zero is a clean run.

A mark out of ten would have thrown most of the run away. A deep sheet carries eight holes, and marking seven of them plus one stray would score the same as marking nothing at all.

Those are not the same performance. The gap between them is where most people who find this hard actually sit, so the score keeps it.

The card splits your number into the two ways of being wrong. One is holes you missed, the other is holes you added that were never there.

They add up to the headline, so you can see which way you go wrong. Missing holes usually means you stopped unfolding early. Adding them usually means you mirrored a fold that was not there.

There is one more thing the format buys. The sheet has sixteen squares to mark, so nothing on this page can be guessed. A version that offered five finished pictures would put a floor of twenty per cent under everybody.

Nothing here is hard to see. The pictures are small and plain, and no square is hidden behind another one.

What runs out is the bookkeeping. Three folds put eight sheets of paper under the punch, and eight positions to work out on the way back.

Most people undo the folds one at a time, in the reverse order to the one they were made in. Each undo doubles what you are carrying.

So the task is really a question about how much you can hold while doing something else to it. The same shortage turns up when people hold digits, and the sizes are in the same range: see what counts as a good digit span.

The mechanism behind both is described in how working memory works, which is worth reading before deciding that a low score here means anything wider.

There is a second folding task that gets called by the same name, and it measures something else.

It shows a flat cross of six squares, the shape a cube makes when it is opened out. Two of its edges carry an arrow, and the question is whether the arrows meet once the cross folds back up.

Shepard and Feng timed that task in 1972, and the answer took longer the more squares had to come along in the fold.

Timing an answer that way is how the turning task next door is measured too. The line it produces is in mental rotation accuracy by angle.

That question has two answers, yes and no, and half of them can be had by tossing a coin. This one has a pattern of holes, and a pattern can be marked square by square and scored the same way.

So the punched holes version is the one on this page. If you came looking for the cube net task, one form of it is an item on the mixed spatial paper. Four sheets are offered there and exactly one of them closes.

Most pages on this site put your result on a curve and tell you where it sits. This one shows the number on its own, and the reason is worth stating.

The paper test this comes from is VZ-2, from a 1976 manual of research tests. It gives twenty items with five finished pictures each, under a three minute clock, and it is scored as a mark out of twenty.

Its norm tables exist. They describe a different answer format, a different item count and a scoring rule with a correction for guessing built into it.

A mark on that test is not an error count on this one. Stretching one to look like the other would be an invented figure wearing a real citation, so the page says there is no comparison yet.

That changes once enough runs have been recorded here. The bar appears when a cohort reaches a thousand results, and not one run before that.