Abstract reasoning test: matrices and patterns

An abstract reasoning test leaves out what you might have learned at school. There are no words to know and nothing to calculate, just shapes that change according to rules you have to discover. If you have been sent Matrigma or the Alva Labs logic test by a Swedish employer, this is the format you will meet. The same format also goes by the names matrix reasoning test and pattern recognition test.

What abstract reasoning means in a test

Alva Labs says its logic test measures abstract reasoning and problem solving: how efficiently you find patterns, draw logical conclusions and deal with new information. The Psychometrics Centre at the University of Cambridge describes Raven’s matrices in similar terms, as the ability to perceive new patterns and relationships.

In practice, you work with forms, symbols and rules instead of facts. Alva says you need no math formulas or language skills beyond understanding the instructions. Assessio describes Matrigma’s tasks as non-verbal and says they measure logical and spatial ability. Matrigma is available in 14 languages, but that only affects the instructions, since the tasks themselves are figures.

Matrix reasoning tests

A matrix task is a 3×3 grid of figures with one cell empty. You look at how the figures change across the rows and down the columns, then choose the piece that completes the grid. In the Alva Labs logic test each task has six options.

The two matrix tests you are most likely to meet differ in pace. Adaptive Matrigma takes 12 minutes with one minute per task. The Alva Labs logic test has 20 scored tasks with two minutes each, and you cannot skip a task or go back. Both adjust the difficulty to your answers, so you will reach tasks that feel too hard whatever your level. The Matrigma and Alva Labs logic test pages cover each test in detail.

Raven’s Progressive Matrices

Raven’s Progressive Matrices is the classic matrix test. It was developed in the UK and first published in 1938. Each item is a geometric pattern with a piece missing, and you choose the piece that completes it from six to eight options. The items get harder as you go, which is where “progressive” comes from, and the standard version has 60 items.

Raven’s is non-verbal and is used to measure general ability and abstract reasoning. Pearson publishes an updated version, Raven’s 2 from 2018, which can be taken on paper or digitally and gives raw scores and percentiles against the population. If you have seen Raven’s, you will recognize the format in Matrigma and Alva Labs, although those are separate tests with their own items, time limits and norm groups.

Pattern recognition beyond matrices

Some abstract tasks use other layouts. In Aon’s scales ix you find the rule in a pattern and pick the one item out of nine that does not fit. In scales cls you work out the rules in a few groups of images and sort new images into the right group. SHL’s interactive inductive tasks have you tap cells in an empty grid to draw the next picture in a sequence.

Test publishers often call these tasks inductive reasoning, because you go from examples to a rule. The inductive reasoning test guide covers that side, including number series, which test the same skill with numbers.

What changes from cell to cell

Most figure tasks are built from a few kinds of change: how many shapes there are, where they sit, which way they point, how they are filled, how big they are, and whether two cells are laid on top of each other to make the third. Easy tasks use one of these. Harder ones combine two or three, or hide a rule in how the figures overlap.

A typical example: in each row the number of dots goes up by one from left to right, while the shape changes fill down each column. You have to hold both rules at once and find the option that satisfies both. Name one rule, check it on every row, then look for the next. Any option that breaks a confirmed rule can be crossed out.

How matrix tests are scored

You are compared with a norm group, not with the other applicants. Matrigma gives a score from 0 to 10 against the working-age population, and about 68 percent land in the middle range. Alva Labs gives a score from 1 to 10 against working adults, and 5 and 6 are the most common results. Because Alva’s test is adaptive, two people with the same number of correct answers can get different scores.

How to practice

Assessio says problem-solving ability is largely stable and that practice makes you more comfortable with the format rather than smarter. Research points the same way: in a meta-analysis of 50 studies, scores rose slightly on a second attempt, most when the same type of test came back. Practice still pays off, because you stop losing points to an unfamiliar format and a ticking clock.

Learn the kinds of change listed above, practice with the time limit of your test, and read the explanation for every task you miss. Start with our free logic test: 12 matrix tasks, one minute each, with new tasks every time. Closer to the real thing, there are full practice tests for adaptive Matrigma and the Alva Labs logic test. For how matrices fit with the other test types employers use, see the logical reasoning test guide. Our tasks are our own and imitate the task type; we are not connected to Assessio or Alva Labs.