Matrix reasoning test: method

What it measures

Non-verbal abstract (fluid) reasoning: inducing the rules that vary across rows and columns of a 3 x 3 matrix and applying them to find the missing cell.

Procedure

Each run draws 12 items at random from the active bank of Sandia matrices (758 items; items no one solved in the original norming were removed). Each item shows a 3 x 3 problem matrix with the last cell missing and a 2 x 4 panel of 8 answer options. The player taps or clicks one option; it is highlighted for 200 ms and the next item appears. There is no going back, no practice and no time limit. Answer keys never leave the server.

Parameters

Scoring

Primary metric: number correct out of 12; accuracy = correct / 12 is also reported. A response to an unknown item counts as wrong. The score is not weighted by item difficulty and no IQ, percentile or norm is shown.

Valid-run rules

A repeated item in one run makes the payload invalid. Leaving the tab or switching windows during the run stops it, and nothing is scored. Every scored run is stored with its input device class (touch, mouse, keyboard), page language, scorer version and runner-config hash. Quality flags (for example many runs from one connection within an hour, a missing visitor id, an automated browser, impossibly fast answers) and earlier attempts by the same visitor, account or connection are recorded with the row. Such runs are kept, but they are left out of any comparison with other players. Flags never change the player's own result.

Differences from the classic task

Raven's Progressive Matrices use a fixed set of items in increasing difficulty with published norms. Here each run is a random draw of 12 items in random order from a larger bank, so the difficulty differs between runs. The items are the Sandia matrices, not Raven's items. There is no time limit and no norm conversion.

Limitations

Twelve items give a rough score, and because each run has a different mix of easy and hard items, raw scores are not strictly comparable between runs or people. Repeated runs can include items seen before. Screen size affects how clearly the details are seen. Unsupervised setting. Not an IQ test and no clinical use.

Languages and review status

Language-independent stimuli; instructions and feedback in all site languages.