Skip to main content
Back to Blogs
Provider Guide7 min read

The Alva Labs Logic Test: 20 Tasks, a 1 to 10 Score and How to Prepare

What the Alva Labs logic and personality tests contain, how Alva turns 20 matrix tasks into a 1 to 10 score, how it differs from Matrigma, and how to prepare.

20
matrix tasks, one at a time
2 min
per task, no overall clock
1 to 10
result, shown the moment you finish
24 months
your result can be reused

If your invitation came from invitation@alvalabs.io with a link to app.alvalabs.io, you have been sent Alva Labs, a Stockholm hiring platform used by Academic Work, SEB and a long list of Nordic employers. You will usually see two cards on your dashboard: a Logic test that takes 15 to 20 minutes and a Personality test that takes about 15. Swedish employers call the first a logiktest, Norwegian ones a logikktest.

Both are adaptive, both are short, and both report on the same 1 to 10 scale. Unusually, Alva publishes exactly how the logic score is calculated, so there is no need to guess.

What the logic test looks like

Each task
A 3x3 grid of figures with the bottom right square marked by a question mark, and six options. Exactly one completes the grid.
Length
20 scored tasks after two unscored practice tasks. Most people finish in 15 to 20 minutes; the maximum is 40.
Clock
Two minutes on each task and nothing overall. A notice appears at 30 seconds. When time runs out the task closes and counts as wrong.
Navigation
Select an option, then Submit. No skipping, no going back and no review. You can pause a limited number of times.
Marking
Nothing extra is deducted for a wrong answer, but an unanswered task is treated the same as an incorrect one. Speed earns nothing at all.

The rules read along rows and down columns and never diagonally, and a row and a column can follow different rules. Alva's wrong options are built from the grid's own pieces: two figures laid over each other, or the right shape with the wrong fill. So the habit that pays is predicting the missing square in words before you look at the options.

How the test picks your tasks

1
Three warm-up tasks
The first is drawn from the easiest tasks in the bank and the next two step up, whatever you answer. Everyone gets the same easy-then-harder start.
2
An estimate after every answer
From task four, Alva updates its estimate of your ability after each answer, weighting each task by how hard it is.
3
A random pick from the ten best-fitting tasks
The next task is drawn at random from the ten unused tasks that would tell Alva most about you at that estimate. That is why two candidates never see the same paper.

The consequence is that you should expect to get plenty wrong. The test hunts for the level where you get some right and some wrong, and Alva says plainly that two people with the same number correct can end on different scores, because the harder tasks count for more.

How the 1 to 10 score is worked out

Alva's manual gives the formula: your ability estimate is put on a standard ten scale with score = 2 × ability + 5.5, rounded to a whole number. The middle of the scale is 5.5 and each point is half a standard deviation. Alva uses no norm group; the scale is anchored to working adults through its calibration against Raven's matrices.

1 to 2
Below average. Roughly the 1st to 6th percentile.
3 to 4
Slightly below average. Roughly the 7th to 30th.
5 to 6
Average. Roughly the 31st to 69th. The most common results.
7 to 8
Slightly above average. Roughly the 70th to 93rd.
9 to 10
Above average. Roughly the 94th to 99th.

Alva's recruiter guide uses a simpler cut: 1 to 3 below average, 4 to 7 average, 8 to 10 above. Most people score between 4 and 7, and Alva's own applicants average about 6.3, so a 6 is the typical applicant result. There is no pass mark: each employer sets a range for the role, and a strong personality match can make up for part of a lower logic result. People who retake the test rarely move more than a point.

Alva is not Matrigma, even though it looks like it

Swedish candidates often meet both, and the two share a format: a 3x3 grid and six options. Everything around the format differs, and practising one on the other's clock will pace you wrongly.

Alva Logic Test
Alva Labs
Two minutes per task, no overall clock.
Always 20 scored tasks.
Scored 1 to 10 on a sten scale, mean 5.5.
You see your result immediately.
Adaptive Matrigma
Assessio
Sixty seconds per item and 12 minutes overall.
As many items as the 12 minutes allow, at least 12.
Scored 0 to 10 on a C-scale, mean 5.
Often only a level, if the employer turns the report on.

For the Assessio side, see our guide to Matrigma and the Assessio names.

The personality test

A five-factor questionnaire: Agreeableness, Conscientiousness, Extraversion, Emotional Stability and Openness to Experience, each split into three facets. You rate short statements about yourself at work from 1 to 5, one at a time, with no clock and a Skip question button that carries no penalty. It adapts too: each facet stops once your answers place it clearly, so most sittings run to about 120 statements, and more if your answers pull in different directions.

Each factor and facet comes back on the same 1 to 10 scale with a named description at each end, such as Reserved to Outgoing, and a middle result reads "Neither Reserved nor Outgoing". Alva is clear that neither end of any scale is better. The employer compares your profile with one set for the role, which usually covers only some facets.

Your result follows your email
Results under 24 months old can be shared with a new employer, and once you have shared with a company, your results go to it again automatically. You cannot retake the tests for that company yourself; only the company can allow it. An employer can also send a supervised validation test of 20 new tasks and compare the two scores side by side. Treat the first sitting as the one that counts.

How to prepare

Alva's fair play policy explicitly allows practising on similar tests beforehand, so preparation is the sanctioned route. What to practise:

  • Full 20-task runs in one sitting, somewhere quiet, so concentration holds through the second half.
  • Naming the rule along the rows and the rule down the columns in words before looking at the options.
  • Submitting a best guess at the 30 second notice rather than letting a task close blank.
  • Taking the three warm-up tasks slowly: they are the easiest you will see.
  • For the personality test, reading every statement to the end, since about a third are worded the opposite way round.

Our Alva Labs practice tests follow Alva's published format and algorithm: 20 tasks, two minutes each, the same warm-up and task selection, and a result on the same 1 to 10 scale. On the real test, stay in the test window; Alva's proctoring records tab switches, other applications and screenshots, and the employer sees the log.

Ready to Start Practicing?

Apply these strategies with our comprehensive practice platform

Start Practising Free