Alva Labs Practice: the Logic Test and the Personality Test

A 1 to 10 scale with no norm group. Alva turns its estimate of your ability into a whole number from 1 to 10 with a published formula, 2 × ability + 5.5, anchored to working adults rather than to a norm group. Its manual calls 5 and 6 average and 9 and 10 above average, roughly the top 6%. Most people score between 4 and 7, and Alva’s own applicants average about 6.3. Alva calibrated the scale against Raven’s Standard Progressive Matrices, the classic matrix test, with a correlation of 0.83.

No pass mark. Each employer sets a range for the role and combines your results into a Role Fit. For less complex roles many accept anything inside a broad band, and Alva says a strongly matching personality profile can make up for part of a lower logic result.

Personality with named ends. Every factor and facet is shown on the same 1 to 10 scale with a description at each end, such as Reserved to Outgoing; a middle result reads "Neither Reserved nor Outgoing". You see the same report the employer does.

Assessio’s Adaptive Matrigma looks almost identical: a 3x3 grid and six options. But Matrigma gives you sixty seconds an item inside twelve minutes overall, while Alva gives you two minutes a task and no overall clock. Matrigma reports a C-score from 0 to 10 with a mean of 5; Alva a sten from 1 to 10 with a mean of 5.5. A candidate who carries Matrigma timings across will rush Alva’s tasks for no reason. If you have both, practise each on its own clock; see our Assessio practice for Matrigma.

Phone or computer. Alva supports phones, so you can practise here on either. Academic Work, which sends Alva for every role it fills, recommends sitting the real tests on a computer with a stable connection.

Alva’s own algorithm. Alva publishes its logic test model, and ours follows it: the same three warm-up tasks, an ability estimate after every answer, the next task drawn at random from the ten that fit best, and the same 1 to 10 formula. What we cannot copy is Alva’s calibrated task bank. Our 150 tasks are written to Alva’s published format with difficulties set by design rather than measured on real candidates, so treat your result as an estimate of the real one.

Our own statements. The personality test uses our own 240 statements written in Alva’s style, 16 per facet, chosen adaptively and stopped per facet in the way Alva describes. Alva’s scoring model is calibrated on hundreds of thousands of candidates; ours is a simple model of your answers. Read the profile as a rehearsal, not a prediction.

No practice phase and no pause. The real logic test opens with two unscored practice tasks and allows a limited number of pauses. Ours goes straight to the first scored task and has no pause.

Not affiliated with Alva Labs. CogniPrep is independent. Alva’s fair play policy explicitly allows practising on similar tests before you sit its own.