Candidate Preparation Guides Without Compromising Assessment Integrity

Two applicants receive invitations to the same online assessment. One has taken several similar tests and knows how to move between questions, interpret the timer and read the instructions. The other is encountering an online skills test for the first time. They may have comparable job-related ability, yet the second applicant could waste valuable time simply learning how the interface works.

Employers can reduce that avoidable disadvantage by giving candidates clear preparation information. The challenge is to help everyone understand the format, rules and technology without revealing the actual questions or coaching only selected people toward the answer key.

A good preparation guide should make the assessment easier to understand, not easier to game. This article shows employers how to create that balance, with examples, an approval process, a sample preparation pack and ways to evaluate whether the guidance is working.

Preparation and security are not opposites

Test security protects the integrity of the evidence collected. Candidate preparation helps people demonstrate the skills the assessment is intended to measure. Both goals matter.

An assessment of numerical interpretation should not primarily measure whether someone can find the Next button. A work-sample task should not reward people who happen to know an undisclosed file-naming convention. A timed test should not surprise candidates with a duration that was never communicated.

At the same time, an employer should not circulate live answer keys, repeat secure questions in promotional materials or offer private coaching to favoured applicants. That would create an uneven opportunity and make the resulting scores harder to interpret.

The useful distinction is between familiarisation and disclosure. Familiarisation explains the experience and may provide representative, non-secure examples. Disclosure reveals protected content or scoring secrets in ways that undermine the intended measurement.

What employers should explain before the test

A candidate preparation pack does not need to be long. It should answer practical questions that a reasonable applicant would otherwise have to guess.

Do not promise a feature that the actual platform does not support. If the system does not allow a candidate to resume after losing connectivity, the guide must not say that progress will always be saved.

The difference between a sample question and a leaked question

A sample question is created or approved for public preparation. It illustrates the style and skills involved without compromising the secure operational pool. A leaked question is content intended to remain confidential, whether disclosed deliberately or accidentally.

The safest practice is to maintain a separate, clearly labelled public sample set. Do not simply copy a few live questions and assume that removing the answers makes them safe. Candidates can still circulate the prompts, reconstruct answers or memorise content that may appear in the real assessment.

If the organisation changes its operational questions regularly, it should still keep the public sample set separate. The preparation material should be reviewed whenever the test format changes, but that does not require exposing new secure items.

A fictional sample question that illustrates the format

An employer assessing basic numerical interpretation could use this non-secure example created only for preparation:

A service team received 80 requests on Monday and 100 on Tuesday. By what percentage did the number of requests increase, using Monday as the base? A. 20% B. 25% C. 80% D. 125%

The correct answer is B: 25%, because the increase is 20 and 20 ÷ 80 × 100 = 25%. The example demonstrates how a percentage-change question is presented. It does not imply that the actual assessment will contain this question or that the real test measures only percentages.

Public examples should use fictional data, have independently checked answer keys and be sufficiently different from protected questions. If the sample is substantially easier than the actual assessment, say that it illustrates format rather than predicting difficulty.

What to include in a 40-question assessment preparation pack

Suppose an employer uses a 40-question online skills assessment with a 45-minute time limit. A preparation pack could contain a short introduction, a description of the assessed domains, three to six public sample questions, a navigation demonstration, technical requirements and support information.

The arithmetic average is 45 ÷ 40 = 1.125 minutes, or 67.5 seconds per question. But that is not an instruction to spend exactly 67.5 seconds on every item. Some questions may require more reading or calculation than others. The employer should explain whether candidates can skip and return to items, rather than advising a strategy that the interface does not allow.

If a test contains several skill areas, the preparation guide should identify them in general terms. For example: “The assessment includes numerical interpretation, reading comprehension and workplace judgement.” It should not publish the exact live blueprint, answer distribution or item sequence if doing so would create a material security risk. The amount of detail disclosed should be consistent with fair candidate information and the assessment's design.

A short optional interface demonstration can be especially valuable for people unfamiliar with online testing. Ideally it should use the same controls as the live experience without counting toward the hiring score.

Build a preparation pack with four layers

Layer 1 — Essential logistics. Include date, start and end window, expected duration, login steps, supported devices and the help route. Write this in plain language that can be read quickly on a phone.

Layer 2 — Assessment overview. Explain the job-related purpose, format and rules. State what candidates should do if they are unsure about a question, whether unanswered items can be revisited and what happens after submission.

Layer 3 — Non-secure familiarisation. Offer representative sample items, a practice interface or a short walkthrough. Clearly separate practice scores from actual selection scores.

Layer 4 — Fairness and privacy. Explain how to request accommodations, report technical problems and find the privacy notice. Make clear that requests are handled confidentially and do not automatically disclose sensitive details to assessors.

These layers can be delivered as one accessible webpage, email or downloadable document. The delivery method matters less than consistent access and accurate information.

Why equal access to preparation matters

Preparation can improve fairness only if the opportunity is available to everyone who needs it. A recruiter who privately sends sample materials to one applicant but not others may create an avoidable advantage. Similarly, placing preparation behind an unrelated paid course could disadvantage candidates who cannot afford it.

A sensible default is to provide the same official pack to all invited candidates at the same stage, sufficiently in advance for them to read it. If candidates enter the process at different times, make sure the notice period remains reasonable.

Candidates may have different language needs, disabilities, device limitations or internet access. Offer accessible formats and a clear way to request help. A candidate should not need a high-end laptop merely to read the instructions if the actual assessment can be taken on a simpler device.

Not every candidate will use the preparation pack. That is acceptable. The objective is to provide an opportunity to understand the assessment, not to force everyone into a particular study routine.

Do not turn familiarisation into unmeasured coaching

Some preparation advice can change the skills being measured. For example, teaching candidates a specific shortcut that directly solves many secure questions may make scores depend on exposure to the coaching rather than the intended skill.

Before publishing guidance, ask:

1. Does this material explain the test experience, or does it reveal protected content?
2. Could an applicant answer secure questions by memorising this pack?
3. Is the advice available equally to all candidates?
4. Does it privilege familiarity with a particular device, software brand or cultural convention without a job reason?
5. Does the sample resemble the real format closely enough to be helpful without misrepresenting difficulty?
6. Has someone checked the accuracy of every answer and instruction?

The employer may appropriately recommend general practice in job-related skills. But it should distinguish broad learning from “inside information” about the actual assessment. Avoid suggesting that candidates can guarantee a pass by studying a particular set of sample answers.

A preparation-policy decision table

“Usually” is deliberate. An assessment used for learning or low-stakes practice may permit more detailed feedback than a high-stakes recruitment test. Employers should define the purpose and risk before deciding what to release.

Who approves candidate-facing material?

Preparation content should have an owner. A practical review workflow involves four checks:

Content review: The subject-matter expert confirms that sample answers are correct and the skills description is accurate.

Assessment review: The assessment owner confirms that no live questions, protected scoring rules or confidential test forms are disclosed.

Candidate-experience review: HR checks clarity, accessible language, timing, support information and consistent delivery.

Technical review: The platform owner verifies that the instructions match the current login, timer, navigation and submission behaviour.

Keep a version number, approval date and record of what changed. If the platform introduces a new resume feature, update the pack. If a security incident affects a sample item, review whether the operational bank was exposed and follow the incident procedure.

A realistic example of a preparation failure

A fictional employer publishes a practice booklet with ten questions. The booklet is copied from an internal pilot, and six of those questions remain in the live 40-question test. Some candidates receive the booklet through a recruiter; others never see it.

The employer then observes that booklet recipients score higher. It would be unsafe to interpret the difference as proof of greater underlying ability. Exposure to repeated items could have contributed to the result, and access was unequal.

The appropriate response is to investigate the affected items, identify which candidates received the material, review whether scores remain interpretable and consider a consistent, fair remedy. It may be necessary to replace exposed questions or reassess candidates with a suitable form. The organisation should not quietly delete the booklet and pretend the results are unaffected.

The example is illustrative; it does not provide a universal statistical rule for how many exposed items invalidate a test. That judgement depends on the assessment's purpose, content, scoring and evidence.

How to evaluate whether preparation improves the process

An employer can collect feedback without treating preparation as a new selection test. Useful indicators include whether candidates understand instructions, whether support tickets decline, whether fewer people abandon the assessment because of navigation problems and whether the practice interface functions on common devices.

Consider a fictional comparison of technical-support requests:

The observed rate falls by 6 percentage points, and the relative decrease is (15 − 9) ÷ 15 × 100 = 40%. This is a descriptive comparison, not proof that the preparation pack caused the improvement. Candidate groups, devices, platform changes or support processes may also have differed.

Collect optional feedback about clarity and accessibility, investigate incidents and avoid unnecessary collection of sensitive personal information. If an employer wants to make stronger causal claims, it needs a suitable evaluation design and appropriate expertise.

A candidate-friendly preparation message

The following example is illustrative and should be adapted to the actual test configuration:

You have been invited to complete a skills assessment lasting approximately 45 minutes. It contains 40 questions about the skills described in the invitation. Before starting, please review the test instructions and try the optional practice demonstration. The demonstration is not scored for recruitment. The invitation explains permitted materials, navigation and what to do if your connection fails. If you need an accessible arrangement, contact the assessment support team before the test window closes. Your result will be considered as one part of the selection process.

This message avoids promising an automatic retake, a particular passing score or features that have not been confirmed. The employer should add the actual contact method, test window and verified technical requirements.

What about third-party preparation courses?

Candidates may use books, videos, practice websites or coaching services. Employers cannot assume that all external preparation is improper. Learning general numerical reasoning or workplace skills can be legitimate.

The concern arises when external providers claim access to confidential live questions, sell stolen answer keys or encourage impersonation. Employers should publish clear integrity rules and provide official preparation information so candidates are less dependent on rumours. Where credible exposure is reported, investigate without automatically accusing everyone who used a preparation resource.

If the assessment uses an open-book or permitted-AI format, say so explicitly. A blanket “no external resources” rule may be inappropriate for a work sample intended to reflect a job where documentation is routinely used. The permitted-tool policy should follow the construct being measured and apply consistently.

How to handle suspected question sharing

When staff discover a shared file that appears to contain live questions, preserve the evidence securely, restrict further exposure and notify the assessment owner. Determine whether the material is genuinely from the operational bank or merely a generic practice question. Review which forms and decisions may be affected.

Do not publish the suspected leaked content while warning candidates about it. Avoid speculative allegations against individuals. Use the documented incident and retake policies to decide whether replacement forms, score reviews or other remedies are needed.

For a separate guide to monitoring compromised items, see How to Detect Question Exposure and Answer Sharing. This is a proposed internal path requiring verification before import.

A two-week preparation-pack rollout

Days 1–3: Confirm the assessment purpose, format, permitted materials, timer, technical requirements and accessibility process. Identify who owns each piece of information.

Days 4–6: Write plain-language instructions and create non-secure example items. Obtain independent answer-key checks and verify that examples do not overlap with the operational bank.

Days 7–9: Test the pack with people unfamiliar with the platform. Ask them to find the help contact, explain the timer rules and complete the demonstration. Fix confusing steps.

Days 10–12: Complete security, HR and technical approval. Version the document and define how updates will be communicated.

Days 13–14: Distribute the pack consistently to invited candidates. Record the version used and provide a route for corrections or accessibility requests.

A simple process that is maintained is better than an elaborate guide that becomes inaccurate after the next product release.

Frequently asked questions

Should employers provide practice questions before recruitment assessments?

Often yes, if they are non-secure, accurate, representative of the format and equally available to invited candidates. Practice questions should not be copied from a live operational bank.

Does preparation make a test less valid?

Not automatically. Familiarisation may reduce irrelevant confusion. However, coaching on secure answers or content can change what scores reflect. The effect depends on the test and the type of preparation; employers should review the evidence rather than make blanket claims.

How many sample questions should we give?

There is no universal number. Provide enough to illustrate each distinct interaction or question format without turning the pack into a replica of the live test. Three to six examples may be practical for a short mixed-format assessment, but that is an editorial suggestion, not a psychometric standard.

Should candidates be told the pass mark?

The answer depends on the assessment's purpose, fairness policy and security considerations. If a minimum standard is used, it must be justified and consistently applied. Disclosure decisions should not be improvised after results are known.

Can we recommend commercial preparation courses?

Be cautious about endorsements and unequal access. Official preparation information should not depend on a candidate paying an unrelated third party. Do not endorse anyone claiming access to secure questions.

What if candidates use AI tools during preparation?

General learning with AI may be different from using prohibited tools during the live assessment. State the actual live-test rules clearly and design assessments that reflect the job's legitimate tool use. Do not imply that AI-assisted preparation itself proves misconduct.

Can we reuse the same sample questions every year?

Public examples can be reused if they remain accurate and separate from secure content. Review them when the format, skills, scoring rules or accessibility features change.

Is a practice score a prediction of the real test score?

Not unless there is appropriate evidence supporting that interpretation. A short, non-secure demonstration normally exists to teach the format, not to provide a reliable forecast of performance.

Related guides and professional references

Employers may also consult Assessment Instructions That Reduce Avoidable Candidate Errors, Accessible Online Assessments and How to Detect Question Exposure and Answer Sharing. All internal routes are proposed and require CMS slug verification before import.

For broader testing guidance, see the International Test Commission Guidelines on Test Use and the ITC/ATP Guidelines for Technology-Based Assessment, updated July 2025. They address responsible test use and technology-related fairness, accessibility, privacy and security concerns; they do not prescribe one universal preparation pack for every employer.

Conclusion

Candidates should not have to guess how an assessment works. Clear instructions, public examples and an accessible practice experience can reduce unnecessary barriers while preserving the meaning of scores. The employer's responsibility is to prepare everyone for the process without revealing what the secure test is supposed to discover. When preparation content is separate from live items, reviewed carefully and distributed consistently, fairness and integrity can reinforce one another.