Engage Logo

Aptitude Testing

Last verified

Aptitude testing is the use of standardised tests to assess a candidate's reasoning, numerical, verbal or role-specific ability before or during selection. The test is only worth running where its scores relate to performance in the actual role, and where the way it is administered does not exclude people for reasons unconnected to the work.

What the common test types actually measure

TypeMeasuresReasonable use
General reasoningAbstract pattern recognition and problem solvingHigh-volume entry-level intake where prior experience gives no signal
NumericalWorking with figures, ratios and tables under time pressurePayroll, finance, analytics and commercial roles
VerbalReading comprehension and inference in the test languageRoles where written communication in that language is core
Domain or technicalSpecific job knowledge already acquiredLateral hiring where the skill is the requirement
Work samplePerformance on a task resembling the real workAlmost any role where a task can be isolated
Personality inventorySelf-reported dispositionDevelopment conversations, not pass or fail screening

Two things follow from this table. Work samples are usually the strongest signal and the least used, because they take effort to construct. Personality inventories are the most used as a gate and the weakest justification for rejecting anyone, since they measure self-report and are transparently gameable by a candidate who has taken one before.

Does the score predict anything here

Vendors present validity evidence drawn from other populations and other roles. Whether it holds for your roles is an empirical question you can answer cheaply.

  • Score everyone at intake, including people you hire, and keep the scores.
  • After a year, compare test scores against confirmation outcomes, appraisal ratings and early attrition for the same cohort.
  • If the correlation is weak, the test is a cost and a candidate barrier producing no selection benefit, and it should be dropped or replaced.
  • Check cut-offs separately. A test can predict performance across its range while the specific cut-off you apply rejects people who would have performed adequately.
  • Re-check after the role changes. A test validated against a role that has since been reorganised is evidence about a job you no longer have.

The most common finding in practice is that a cut-off was set to reduce interview load rather than to select on ability, and has never been revisited. That is a defensible reason to use a test, but it should be stated as such rather than dressed as prediction.

Where a test becomes a barrier

A test screens on everything it requires, not only on what it intends to measure.

  • Language. A verbal or numerical test written in English measures English reading alongside the target ability. For roles where the work is done in a regional language, that is a barrier with no job relevance, and it disproportionately affects candidates from vernacular-medium education.
  • Familiarity. Candidates from urban schools and coaching-heavy backgrounds have usually seen the format before. The first-time taker is slower for reasons unrelated to ability, which is an argument for practice questions before the timed section.
  • Disability. Candidates with visual, hearing, motor or learning disabilities are entitled to accommodation in the assessment, which can include additional time, a scribe, accessible formats or assistive technology. This has to be offered and arranged in advance, which means asking about requirements at the invitation stage rather than on the day.
  • Device and connectivity. A timed online test taken on a shared phone over an unstable mobile connection is not the same test as one taken on a laptop, and the difference shows up as a score.

Where a group of candidates consistently scores lower and the test is not clearly job-related, the exposure is real. Verify the accommodation obligations under the Rights of Persons with Disabilities Act, 2016 and the guidance issued for written examinations before finalising an assessment process.

Administering it without losing the signal

  • Decide proctoring by stakes. An unproctored screening test at volume will be taken by someone other than the candidate in a minority of cases, which is tolerable at the first filter and not at the final one.
  • Re-establish the score later. A short in-person or proctored re-test at the interview stage, on the same material, deters impersonation more cheaply than proctoring everyone.
  • Publish the retest policy. Whether a candidate can retake a test, after how long, and which score counts should be decided in advance, because it will otherwise be decided case by case in favour of whoever asks.
  • Do not use a proctoring method you cannot justify. Continuous webcam recording, screen capture and behavioural flagging collect a great deal of personal data about candidates, and each capability needs a purpose, a notice and a retention period.
  • Tell candidates what the test is, how long it takes and what happens to the result. Assessment is the stage candidates report as most opaque, and the drop-off there is concentrated among people with other offers.
  • Fix the retention of scores and recordings with the vendor in writing. The vendor processes candidate data on your behalf, and the obligation to the candidate remains yours.

Statutory reference

Act
Rights of Persons with Disabilities Act, 2016
Section
Rights of Persons with Disabilities Act, 2016: non-discrimination in employment, the duty to provide reasonable accommodation, and accessibility of the selection process; accommodations in written assessment such as additional compensatory time, a scribe, and accessible formats, as elaborated in the guidelines issued for the conduct of written examinations for persons with benchmark disabilities. Digital Personal Data Protection Act, 2023 and the rules made under it: notice, purpose limitation, security and erasure in respect of test scores, recordings and proctoring data, and responsibility for an assessment vendor acting as a processor. Code on Wages, 2019: prohibition of discrimination in recruitment for the same work or work of a similar nature
Key limits
Reasonable accommodation in the assessment itself must be offered and arranged in advance, not only in the workplace after hiring. Proctoring data and test scores are personal data with notice and retention obligations, and engaging an assessment vendor does not move those obligations off the employer. Verify the accommodation entitlements, including the compensatory time and scribe provisions in the written examination guidelines, before setting an assessment process. Under notification G.S.R. 843(E) dated 13 November 2025, sections 3 to 17 - the grounds for processing, notice, consent, the general obligations of a data fiduciary and all of the data principal rights - take effect eighteen months from that date, on 13 May 2027. Only the definitions and the Data Protection Board and penalty machinery are in force now. Separately, processing for employment purposes runs on the section 7(i) legitimate use and not on consent, and the erasure duty in section 8(7) yields where retention is necessary for compliance with any law, which is the position for statutory payroll and register retention. Section 20 non-discrimination in employment, the section 20(2) duty to provide reasonable accommodation and a barrier-free environment, the section 20(4) protection of an employee who acquires a disability in service, the section 23 Grievance Redressal Officer and the section 34 four per cent reservation all bind government establishments only. What binds every establishment, private included, is section 21, notifying an equal opportunity policy and registering it with the Chief or State Commissioner, and section 22, maintaining records of persons with disabilities open to inspection. For the private sector section 35 provides government INCENTIVES towards a five per cent workforce target, not a quota. The general prohibition that does reach private conduct is section 3(3), and it carries a defence where the act or omission is a proportionate means of achieving a legitimate aim. Reasonable accommodation is defined in section 2(y) as excluding what would impose a disproportionate or undue burden. Sections 3 to 17, which carry the duties described here, commence on 13 May 2027 under G.S.R. 843(E) of 13 November 2025. In force now are the definitions, the Data Protection Board and the penalty provisions.

Source

Frequently asked questions

What is aptitude testing?

The use of standardised tests to assess a candidate's reasoning, numerical, verbal or role-specific ability during selection. It is worth running where the scores relate to performance in the actual role and where the administration does not exclude people for reasons unconnected to the work.

How do we know whether a test works for our roles?

Score everyone at intake, keep the scores, and after a year compare them against confirmation, appraisal ratings and early attrition for the same cohort. Vendor validity evidence comes from other populations and other roles, and does not answer the question for yours.

Do we have to accommodate candidates with disabilities in a test?

Yes. Accommodation applies to the selection process itself and can include additional time, a scribe, accessible formats or assistive technology. It has to be arranged in advance, which means asking about requirements when the assessment is scheduled rather than on the day.

Is an English aptitude test a problem?

It is if the role is not performed in English. The test then measures English reading alongside the target ability and disadvantages candidates from vernacular-medium education for a reason unconnected to the job.

How do we handle impersonation in remote testing?

Treat the unproctored score as a first filter rather than a final one, and re-establish it with a short proctored or in-person test at the interview stage. That deters impersonation at far lower cost, and with far less candidate data collected, than proctoring everybody.

How Engage keeps assessment records usable

Engage holds test scores against the candidate record with the date and the stage they were taken at, so a cut-off can be checked a year later against confirmation and attrition outcomes for the same cohort. Accommodation requests are captured at invitation rather than on the day, and assessment data carries a retention position rather than sitting with a vendor indefinitely.

See recruitment in Engage
WhatsApp