Reasoning6–8 minBurden: MediumEMA: Low

Cube Folding

A spatial visualisation task: a cube net with a symbol on each square, and four drawn cubes of which exactly one can be folded from it. Every item is generated from a seed.

Spatial visualisationMental foldingSpatial ability
Category
Reasoning
Typical duration
6–8 min
Participant burden
Medium
EMA suitability
Low

Try this task

Configure parameters and run an interactive preview exactly as participants will experience it. No data is recorded.

Configure preview

Adjust parameters below, then start the preview on the right.

Include practice trials

Shown with feedback before the main task

Task parameters

Shared evenly across the levels, easiest first; the 11 cube nets take turns. Default 15, 5 per level (our choice; published net and cube tests use 11–20 items, e.g. Make-A-Dice, Burte et al., 2019).

With feedback; one per level in turn, after a worked example. Default 3 (our choice).

Levels

Level 1: symbols that look the same after any turn, so only which face is where matters. Level 2: symbols with a direction (arrow, L, T, …), so the way each one points after folding matters too. Level 3: as level 2, with the shown faces far apart on the net, so more squares have to be folded along; in mental paper folding, response times rise with the number of squares carried (Shepard & Feng, 1972; Wright et al., 2008).

Default none (our choice: a power measure with RTs). Published paper tests limit the whole block instead (see the next field). An item not answered in time counts as an omission.

Default off (our choice). Published tests limit the whole block: the DAT-5 Space Relations test gives 20 min for 50 items, about 24 s per item (Ramírez-Uclés & Ramírez-Uclés, 2020), which would be 6 min for 15 items here. With a block limit, the items come in order without going back, a countdown is shown, the item on screen when the time is up is kept but not scored, and items never shown are reported as not reached. Practice is not timed.

Empty: each participant gets a new form, and the seed it was drawn from is stored with the result. The same seed gives every participant the same items.

Four cubes per item: the answer, a cube with each of two error types (swapped faces, a symbol turned the wrong way, opposite faces side by side) and one with both, so the answer cannot be found from the cubes alone (facet design; Pallentin, Danner & Rummel, 2023; Hu et al., 2021). About 7 min with these settings. Outputs: correct, answered, accuracy per level and per error type (chance 50 % per type), median RT of correct answers, and with a block limit the items not reached; each trial stores the net, the cubes and the errors of the chosen cube.

6–8 minBurden: MediumEMA: Low

This is a researcher preview. No participant data is recorded.

Simulated participant view

9:41

Cube Folding

A spatial visualisation task: a cube net with a symbol on each square, and four drawn cubes of which exactly one can be folded from it. Every item is generated from a seed.

No data is recorded

Participant experience on smartphone

The net (one of the 11 cube nets, in any of its rotations and reflections) at the top; below it four cubes in isometric view, each showing three faces, in a 2 × 2 grid, tapped to answer. Symbols are filled shapes in one dark ink on white or light grey faces, mapped onto each face, so nothing depends on colour.

When to use

For a spatial visualisation measure with many parallel forms: items are generated, so sessions can use new items (no seed) or one fixed form (a set seed). Suited to baseline and occasional sessions, and to studies of the cube and net items used in aptitude and selection tests.

When not to use

Not a normed spatial ability test: the items follow published task formats and distractor principles, but have no norms or IRT parameters yet. Not a quick EMA probe. For 2-D rotation, use Mental Rotation; for figural reasoning, Figure Series.

How to use in a study

The defaults give 15 items, 5 per level, levels 1 to 3, easiest first, without a time limit. Leave the seed empty to give each participant and each session new items; set a seed to give everyone the same form. A worked example and one practice item per level with feedback come first. To approximate the time pressure of published paper tests, set a limit per item: their block limits work out at about 24–27 s per item (DAT-5 Space Relations; IST 2000 R cubes).

Researcher-configurable parameters

  • Number of items: default 15, 5 per level (our choice; published net and cube tests use 11 to 20 items per form or block)
  • Levels: 1 plain symbols (only which face is where matters), 2 symbols with a direction (the direction after folding matters too), 3 symbols with a direction on faces that lie far apart on the net
  • Time limit per item: default none; about 24 s per item matches the DAT-5 Space Relations block limit (50 items in 20 min)
  • Time limit for the whole block: default off; 6 min for 15 items matches the DAT-5 Space Relations pace (50 items in 20 min)
  • Seed: empty (new items per run; the seed is stored with the result) or a number (the same items for everyone)
  • Practice items with feedback: default 3, one per level (our choice; published tests give one or two worked examples)

Outputs collected

  • Correct answers and items answered
  • Accuracy per level
  • Accuracy per distractor type (swapped faces, symbol direction, opposite faces): among the items whose options test that type, the share where the chosen cube does not make that error (chance 50 %)
  • Median RT of correct answers
  • Every trial: level, net and its placement, the symbols with their turns on the net, the distance between the shown faces on the net, the four cubes, the type of each wrong cube, the correct cube, the response and its errors, touch-down RT, time-outs
  • The seed, from which every item of the run can be rebuilt

Interpretation notes

Wrong cubes are of three types: two shown symbols swapped (the mirror image of a possible view), one symbol turned the wrong way, or two faces that are opposite on the cube shown side by side. When every wrong option is the answer with one change, the answer is the option that shares the most with the others and can be found without the net (Hu et al., 2021). Here the four cubes are a 2 × 2 facet design instead (Pallentin, Danner & Rummel, 2023): the answer, one cube with each of two error types, and one with both. Each symbol in each place appears in two of the four cubes, so choosing by shared features, by the majority, as the odd one out, or by which cubes could show the same cube is right at chance (1 in 4) on generated items. Level 1 tests swapped and opposite faces; levels 2 and 3 always test the direction of the symbols, together with one of the other two. Level 3 shows faces that lie far apart on the net (5 or more squares between the three pairs, against at most 3 at levels 1 and 2), so more squares have to be folded along: in mental paper folding, response times rise with the number of squares carried (Shepard & Feng, 1972; Wright et al., 2008). Accuracy per distractor type has a chance level of 50 %, accuracy per level and overall of 25 %. RTs are timed from the drawn item to the touch-down, but this timing has not been validated on real devices. The items have no norms; compare groups or sessions within a study.

Scientific evidence

  • In mental paper folding, the time to decide whether two marked edges of a net meet grows with the number of squares that have to be carried along in the imagined folds (Shepard & Feng, 1972; Milivojevic et al., 2003).
  • In a computerised net-and-cube task, response times were linearly related to the number of squares carried (Wright et al., 2008).
  • Make-A-Dice uses all 11 cube nets, one per item; the nets differ in difficulty (Burte et al., 2019).
  • Published net-and-solid tests show a pattern and four drawn solids (DAT Space Relations: 50 items in 20 min, about 24 s per item; Ramírez-Uclés & Ramírez-Uclés, 2020); the ETS Kit's Surface Development test pairs a net with the solid it folds into and asks which edges correspond (Ekstrom et al., 1976).
  • Answer sets that can be solved without the problem inflate scores; a balanced option set removes that (Hu et al., 2021; Pallentin, Danner & Rummel, 2023).