Cognitive ability tests and general mental ability

A cognitive ability test is the part of a job application that checks how you think rather than what you know. You may see it called an aptitude test, a logic test or a general mental ability test. The tasks differ between publishers, but they aim at the same thing: how quickly and accurately you solve problems you haven’t seen before. This guide covers what that ability is, why employers measure it and what the score can and can’t tell them.

What general mental ability is

Intelligence has been studied since the 19th century, and researchers describe several kinds of it, such as numerical, verbal and spatial ability. In the early 20th century Charles Spearman showed that these usually converge in one common factor, called the g-factor or general mental ability (GMA).

Assessio, the publisher of Matrigma, describes GMA as a mix of two parts. Fluid intelligence is the ability to reason, form concepts and solve problems with unfamiliar information. Crystallized intelligence is your store of knowledge and experience, and it can keep growing into old age.

Hiring tests lean toward the fluid part. Alva Labs says its matrix format minimizes the role of previous knowledge and experience. Assessio is careful to say that Matrigma measures one part of intelligence, problem-solving ability, and isn’t a full intelligence test.

Why employers use cognitive ability tests

Employers use these tests because the result shows something a CV doesn’t. Assessio writes that problem-solving ability can’t be identified through a CV or evaluated reliably in a traditional interview, and that some employers use the test to decide which candidates to invite to an interview.

The publishers also say the tests predict how people will do in the job. Alva Labs says logic tests have been shown to be the single assessment tool that best predicts future job performance, across a large variety of roles and industries. Assessio says Matrigma predicts an individual’s future job performance with high confidence. Both tie the effect to the job itself: according to Alva, the more complex the role, the more logical ability matters, and Assessio advises employers to set their requirements by how complex the role is.

These are the publishers’ own descriptions, and Assessio adds a caveat. A high score means you learn and solve problems faster, which doesn’t necessarily make you better at something, and a person with a lower score can compensate with determination or structure. Employers often combine the ability test with other parts of the process. Nacka Municipality, for example, recommends using Matrigma together with the MAP personality test. The guide to psychometric tests shows how the tests fit into a Swedish hiring process.

What the tasks look like

The label covers several task types, and the mix depends on the publisher:

• Matrices: a 3×3 grid of shapes with one cell missing. Matrigma and Alva Labs use them, and the abstract reasoning test guide covers them along with Raven’s Progressive Matrices.
• Deductive tasks: work out what must follow from given rules. See the deductive reasoning test guide.
• Inductive tasks: find the rule behind a pattern or series. See the inductive reasoning test guide.
• Numerical and verbal tasks: tables, charts and short texts. See the guides to the numerical reasoning test and the verbal reasoning test.

SHL Verify Interactive G+ combines three of these. It measures deductive, inductive and numerical reasoning and gives you a score for each, plus an overall score. More on the SHL test page.

If you are applying in Sweden in English, the matrix format has a practical upside. Alva Labs says its non-verbal test is less sensitive to differing levels of language proficiency and needs no math formulas or language skills beyond understanding the instructions. Verbal tasks are different, since reading is what they test.

How it differs from an IQ test

A cognitive ability test and an IQ test measure similar abilities, but a test in recruitment doesn’t give you an IQ score. You get a result on a short scale, compared with a stated group. Matrigma scores from 0 to 10 against the working-age population, and about 68 percent land in the average range. Assessio points out that because the norm group is the working-age population, the result doesn’t tell you how you compare with everyone. Alva Labs scores from 1 to 10 against working adults, where 5 and 6 are the most common results.

Both tests are adaptive in their usual versions, so the difficulty follows your answers and you get harder tasks when things go well. That is why two people with the same number of correct answers can get different results at Alva Labs.

Our free IQ test mixes matrices, number series, word meanings and cubes. It gives you your number of correct answers and never an IQ score, because we have no norm group.

Can practice raise your score?

Somewhat. A meta-analysis of 50 studies with more than 134,000 participants found that scores rose by 0.26 standard deviations on average on a second attempt. The gain was larger after coaching and when the same form of the test came back.

Assessio says problem-solving ability is largely innate and stable in adults and doesn’t improve with education or training. The two findings fit together. Practice makes the format and the clock familiar, so you stop losing points to them, while the ability underneath stays about where it was.

Start with the publishers’ own material: SHL and Aon both have practice tasks. Then take our free IQ test for a mix of task types, or the free logic test if your test is all matrices. When you know which test you will take, try a full practice test with the same structure and time limit. Everything is free and needs no account. Our tasks are our own and imitate the task types, and we can’t promise any result.