Cognitive Ability Test

What a cognitive ability test is

A cognitive ability test is an instrument administered to candidates to measure general reasoning capacities: verbal reasoning, quantitative reasoning, abstract reasoning and speed of comprehension.

Its distinguishing property is that it measures the handling of new information rather than an accumulated stock of professional knowledge. Its items are written so that a candidate needs no background in any particular field in order to answer them.

What it measures, and what it does not

• It measures reasoning from material presented to the candidate inside the question itself.
• It does not measure knowledge of the job, which belongs to a knowledge test or to a practical exercise built on the content of the occupation.
• It does not measure conduct at work, dealings with other people, or motivation.
• It does not measure what the candidate has actually accomplished, which is the territory of questions about specific events and of reference checking.

That places it below the level of a competency. Competencies combine a skill, the knowledge supporting it and the behaviour by which it shows up at work, described in observable terms. A cognitive ability test reaches none of that. It measures a capacity underneath a competency rather than a competency as a framework describes one.

Where it sits in the run of a hiring process

It is usually placed after the first screen and before interviewing, because it can be administered to large numbers at a cost per candidate that barely moves with volume, which reduces how many people reach the stages that are expensive in time. It is used most in three situations: entry level appointments and graduate programmes, where candidates have no professional record to compare; high volume competitions, where interviewing everybody is not possible; and roles whose content turns over quickly, where the capacity to learn matters more than today’s stock of knowledge.

The cut score is a decision in the organisation, not a property of the test

The test produces scores. The organisation decides where the line sits that a candidate has to clear. The effect of that decision is numerical, and it can be computed before it is taken, from the score distribution of a previous cohort.

Take a cohort of 200 candidates whose scores fell out as follows: 24 candidates scored 80 or above; 46 scored between 70 and 79; 58 scored between 60 and 69; 44 scored between 50 and 59; and 28 scored below 50. Those five bands account for all 200 candidates and no candidate falls in more than one.

Put the line at 60 and 128 candidates clear it, which is 64.0 percent of the 200 in the cohort. Put it at 70 and 70 candidates clear it, which is 35.0 percent of the same 200. Ten points in the position of the line removed the 58 candidates in the 60 to 69 band. Those 58 are 29.0 percent of the 200 in the cohort, and 45.3 percent of the 128 who had cleared the lower line.

Nothing about the test, its items or its marking changed between the two cases. A number in the organisation’s policy changed. This is why it is not accurate to say that the test excluded 130 candidates at the higher line. The test ranked them. The line excluded them.

One further thing belongs in the same calculation: the cost of an error here is not symmetric. A candidate who clears the line enters later stages that give the process opportunities to correct its estimate of them. A candidate who does not clear it leaves and is never revisited. So the higher the line goes, the more people exit on a judgement the process has no mechanism for reviewing, and that cost appears in no report, because the people bearing it are not in the data.

Where it sits beside the other selection instruments

A structured interview is a conversation with fixed questions and a defined scale; this is a written measure marked against a single key. They answer different questions, so neither substitutes for the other. Behavioral interview questions ask about situations that actually occurred, so their material is the candidate’s past, while this instrument’s material is live performance on a problem they have not seen before. A work sample test hands over a piece of job output and argues from its content; this argues from nothing in the job’s content at all, which is exactly why it transfers across roles and exactly why it says less about any one of them. And a situational judgement test rests on knowledge of the work context, since what counts as the best option in an administrative situation is not what counts in a field one, whereas this instrument is written to need no such context.

What spoils its use

• Making it an absolute barrier that excludes a candidate without regard to their other evidence, rather than one input into the decision.
• Administering it to some candidates and not others for the same role, which makes the comparison between the tested and the untested meaningless.
• Varying the conditions: open time for one candidate and a limit for another, or supervised administration in one case and unsupervised in another.
• Reusing items that have circulated, at which point the instrument measures prior exposure rather than capacity.
• Using it to appraise a serving employee, which is an entirely different subject: an employee’s performance is measured by what they delivered in their role, not by a general capacity measured before they were hired.
• Overlooking the arrangements a candidate may need in order to sit it.

The result is personal data

A candidate’s score, their name and the record of their answers are personal data under the Personal Data Protection Law (نظام حماية البيانات الشخصية), issued by Royal Decree M/19 of 9/2/1443H and amended by Royal Decree M/148 of 5/9/1444H. Article 2(1) of that law applies it to any processing of personal data relating to individuals taking place in the Kingdom by any means, and Article 1(5) of the same law treats collection, recording, storage, retrieval, use, disclosure, sharing and destruction as processing. A candidate is an individual, and the law’s scope does not turn on an employment relationship existing.

Among the duties it establishes, these bear directly on this instrument:

• Purpose. Article 10 of the Personal Data Protection Law permits processing only to achieve the purpose the data was collected for, subject to the exceptions set out in it. Article 11(1) of the same law requires the purpose to relate directly to the controller’s own purposes and not to conflict with any provision established in law. Article 13(2) of the Personal Data Protection Law requires the data subject, on direct collection, to be told the purpose and which items are mandatory and which optional.
• Minimum. Article 11(3) of the Personal Data Protection Law requires the content of the data to be adequate and confined to the minimum necessary for the purpose of its collection.
• Protection. Article 19 of that same law requires the controller to take the organisational, administrative and technical measures necessary to safeguard the data, including when it is transferred, in accordance with what the regulations specify.
• Destruction. Article 18 of the Personal Data Protection Law requires destruction once the need has ended, with retention only on a legal basis or for a pending case.
• The data subject’s rights. Article 4 of that same law establishes the right to be informed, the right of access, the right to obtain a copy, the right of correction and the right of destruction.

Where an external supplier administers the test on the organisation’s behalf, the organisation is the controller and the supplier the processor, under Articles 1(18) and 1(19) of the same law. Everything beyond this in the law and its regulations is outside what this page establishes, and its reference is the Saudi Data and Artificial Intelligence Authority (الهيئة السعودية للبيانات والذكاء الاصطناعي) and the text of the law itself.

Occupational fitness is a separate matter whose rules do not transfer here

A selection test run by an employer is regularly confused with the occupational fitness examination, and they differ in their subject, in who conducts them and in what reaches the employer. The regulation on occupational fitness examinations and noncommunicable diseases (لائحة فحوصات اللياقة المهنية والأمراض غير المعدية), issued by Ministerial Decision 33232 of 11/3/1447H, governs a medical examination with its own practitioners, and its Article 17(4) limits what reaches the employer to the final determination alone: fit, fit with restrictions or considerations, or unfit, without disclosing the diagnosis or the medical data.

That wall exists to keep clinical detail away from the employer. Carrying the rules of that regulation across to a selection test the employer administers itself would invert the protection into a route toward what it was built to prevent reaching. So each is read on its own terms: that regulation for its own examination, and a selection test on what is actually asked about it in its own place.

Questions to settle before administration rather than after

• Retaking. May a candidate sit it again? After how long? On the same form or a different one? Which of the two scores counts? An answer improvised at the first request becomes a precedent applied to everybody after them without ever having been decided.
• Administration arrangements. A candidate needing an arrangement in how the test is administered is considered before the day rather than on it, and the approach taken is written down so it does not vary from one case to the next.
• How long results are kept. Fixed in advance and tied to the purpose they were collected for, rather than left open in a file that survives for years.
• Who sees the score. Confining access to those who need it to do their work is good practice the organisation adopts by its own decision, and it is a recommendation here rather than a duty we attribute to any text.
• What the candidate is told. Do they receive their score? In what detail? When? Decided once and applied to everybody.

What is left after the score

The test produces a number, and a number is not a decision. A decision rests on several pieces of evidence read together, and whether it was right is judged later by what shows up in the work. An organisation that administers the instrument for years without ever looking at what became of the people it hired does not know what the instrument bought it, and the ranking it produced each year was never checked against anything.

What this page does not establish

We did not find, in our sources, anything establishing how far these instruments predict performance, nor their stability, nor a numerical ranking among selection instruments, nor their effect on particular groups of candidates. Every proportion and coefficient circulating on those questions is outside our sources and is not reproduced here.

This page also does not endorse any particular instrument or any body issuing one, and it does not establish whether anything in Saudi Arabia regulates the use of selection tests in the private sector. Our sources record a search for such a rule that did not find one, which is a fact about our sourcing rather than about the statute book. Anyone needing an answer there should go to the competent authority and its own published document rather than to a general description of the concept.

The score distribution used in the worked example is assumed in order to show how the calculation runs. It is not the output of any existing instrument and is not attributed to one.

Related terms

• Situational Judgement Test
• Work Sample Test
• Test Validity
• Integrity Testing
• Competencies