Work Sample Tests for Recruitment: Design, Scoring and Practical Examples

A candidate may describe excellent spreadsheet skills in an interview but struggle to reconcile a simple dataset. Another candidate may be less polished in conversation yet complete the task accurately and explain the result clearly. Work sample tests help employers observe performance on tasks resembling real work instead of relying exclusively on claims about ability.

The principle is straightforward: ask candidates to complete a relevant, bounded piece of work and evaluate the output against criteria established before scoring begins. The difficult part is designing a task that represents the job fairly, fits the recruitment stage and yields scores that can be interpreted responsibly.

This guide covers work sample design, assessment blueprints, objective and judgement-based scoring, practical examples, candidate experience, confidentiality and quality assurance. It also explains why realistic tasks still require evidence of reliability, fairness and validity.

What is a work sample test?

A work sample test asks an applicant to demonstrate a defined aspect of job performance through a practical task. Examples include reconciling transactions, diagnosing a software defect, responding to a fictional customer complaint or analysing a short operational report.

It differs from a general aptitude test, which may measure broader underlying abilities, and from an unstructured interview, which usually asks the candidate to explain what they would do rather than produce the work itself.

These categories can overlap. A practical exercise may include a short written explanation, and an assessment centre may contain work samples.

Why employers use work samples

A well-designed work sample can provide direct evidence about whether an applicant can perform an important task. It also creates an opportunity to assess output against consistent standards.

However, a realistic-looking task is not necessarily a good selection instrument. The task may represent only a narrow part of the job, use unfamiliar tools, require excessive unpaid effort or be scored inconsistently.

The aim is representative evidence, not a miniature version of the entire job.

Step 1: Identify critical job tasks

Start with a job analysis. Ask what the person must do, what errors matter and which competencies are necessary on entry rather than expected to develop after training.

For a reporting analyst, essential tasks might include reconciling data, investigating discrepancies and communicating findings. For a customer service officer, they might include interpreting policy, resolving complaints and recording next steps.

Avoid adding difficult tasks simply because they are impressive. Each task should support an explicit job-related inference.

Step 2: Create an assessment blueprint

A blueprint specifies what the task measures, how much evidence is needed, how it will be scored and how long candidates should reasonably take.

Illustrative blueprint: Accounts payable work sample

The timing and weighting are examples, not universal standards. They should be piloted and justified against actual job requirements.

A blueprint should also specify permitted calculators, reference materials, spreadsheet software, accessibility accommodations and whether candidates may ask clarifying questions.

Step 3: Build a realistic but bounded task

Use synthetic or appropriately de-identified materials. Never expose real customer records, confidential employee information, passwords or live production systems to applicants.

The task should have enough context to resemble work but not so much that candidates spend most of their time decoding irrelevant background information.

Example 1: Accounts payable reconciliation

A fictional supplier statement lists three invoices:

Candidate task: Calculate the discrepancy, identify the invoice requiring investigation and draft a two-sentence reconciliation note.

Reference answer: The total difference is ₦250,000 − ₦223,000 = ₦27,000. Invoice INV-102 differs by ₦27,000. The candidate should recommend verifying the underlying invoice and ledger entry rather than silently adjusting either record.

A candidate who identifies the correct difference but immediately changes the ledger without supporting documentation may demonstrate numerical accuracy while showing weaker control judgement.

Example 2: Customer complaint response

Scenario: A customer received an order late. The fictional company policy allows reimbursement of delivery charges for verified delays, while a product refund requires separate eligibility review.

Candidate task: Draft a response acknowledging the issue, stating the available remedy and explaining the next action.

A strong response should interpret the supplied policy accurately, avoid unsupported guarantees and communicate a feasible next step. The scoring guide should reward these behaviours rather than unnecessary length or a particular personal writing style.

Example 3: Data quality investigation

A fictional monthly sales extract contains duplicate transaction IDs, missing dates and one unusually large value.

Candidate task: Identify the quality issues, state which checks should precede reporting and explain why the outlier should be investigated rather than automatically deleted.

A good response distinguishes an unusual observation from a confirmed error. It may recommend validating source records, documenting corrections and reporting unresolved uncertainty.

The test should not require knowledge of a proprietary tool unless that tool is genuinely necessary for the role or candidates are given adequate instructions.

Step 4: Choose the right scoring method

Use objective answer keys for verifiable calculations and structured rubrics for judgement-based outputs. Some work samples need both.

Worked scoring rubric: Accounts payable exercise

Suppose a fictional candidate earns 30, 20, 15, 12 and 8 points respectively. Their total is 85/100.

This score summarises performance on the defined exercise. It does not imply an 85% probability of job success or establish that 85 is an appropriate hiring cutoff.

The scoring manual should include partial-credit examples, treatment of arithmetic carried forward from an earlier mistake and instructions for inconsistent or incomplete responses.

Step 5: Calibrate human assessors

When a response requires judgement, have assessors independently score sample submissions and explain their ratings. Compare results and resolve ambiguous rubric language before live use.

For example, two assessors may disagree about whether a candidate's proposed escalation is sufficiently specific. The rubric should clarify whether naming an owner, timing and documentation is required for full credit.

Formal scoring consistency should be evaluated using methods appropriate to the score type and intended use. A single calibration discussion does not establish reliability.

Step 6: Set defensible time limits and conditions

Time limits can affect what a work sample measures. If speed is essential to the job, a justified timed exercise may be relevant. If accuracy and reasoning matter more, an unnecessarily short limit may introduce irrelevant pressure.

Document:

• The expected task duration and permitted preparation.
• Required equipment and any alternative delivery arrangements.
• Whether internet research, calculators or reference materials are allowed.
• How technical interruptions are handled.
• How accommodations are requested and administered.
• Whether candidates receive the same materials or equivalent forms.

Pilot timing with people who understand the job but are not involved in writing the task. Review whether time pressure prevents candidates from demonstrating the intended competency.

Step 7: Avoid excessive candidate burden

A work sample should be proportionate to the recruitment stage. A brief screening task may be appropriate for a large applicant pool; a more involved simulation may be reserved for shortlisted candidates.

Employers should not disguise unpaid commercial work as an assessment. Avoid asking applicants to prepare usable client deliverables, solve current business problems or create intellectual property that the company intends to exploit without clear agreement and appropriate compensation.

Give candidates enough information about the assessment's purpose, duration, scoring and handling of submitted materials.

Step 8: Review fairness and accessibility

Work samples can be unfair when they require experience with irrelevant software, rely on culturally specific knowledge, use inaccessible file formats or impose conditions unrelated to job performance.

Review whether:

1. The task reflects actual entry requirements.
2. Candidates have comparable instructions and resources.
3. Accessibility needs can be addressed appropriately.
4. Scoring rewards job-relevant evidence rather than presentation polish alone.
5. The exercise contains assumptions unrelated to the role.
6. Outcome differences warrant further investigation.

Fairness does not require ignoring essential competencies. It requires distinguishing those competencies from unnecessary barriers.

Step 9: Validate the interpretation of results

Content relevance is an important starting point, but a work sample's realistic appearance does not prove that it predicts later performance.

A validation plan may examine:

Employers should be cautious about strong predictive claims based on a small pilot or a single hiring cohort.

For the broader evidence framework, see How to Validate Pre-Employment Assessment Tests.

Step 10: Protect test security and maintain versions

Repeated use can compromise a work sample when answer keys circulate or candidates share complete submissions. Maintain a version register, restrict answer-key access and periodically review item exposure.

Alternative versions should test comparable competencies and difficulty. Simply changing numbers does not guarantee equivalent scores, especially when the new figures alter task complexity.

Retain scoring rubrics and version identifiers so decisions can be traced to the instrument actually administered.

Combining work samples with other methods

A work sample can be one component of a structured hiring process. For example, an employer might use a practical test to assess technical accuracy and a structured interview to explore communication and judgement.

Consider an illustrative scorecard:

The calculation is (84 × 0.60) + (75 × 0.40) = 80.4%. The employer must still justify the weights and consider whether essential technical failures should be treated separately rather than averaged away.

For a detailed comparison, see Structured Interviews vs Skills Assessments.

Common work sample design mistakes

Testing too much at once. A sprawling task makes it difficult to identify which competency explains a score.

Using real confidential information. Candidate exercises should use safe, synthetic or properly de-identified materials.

Scoring by overall impression. Clear answer keys and rubrics improve transparency.

Making the exercise excessively long. Burden can increase without corresponding evidence of selection value.

Confusing polish with competence. Presentation should be scored only when relevant to the job.

Treating all versions as equivalent. Changes in content or difficulty may affect score comparability.

Using arbitrary pass marks. A familiar percentage is not a defensible minimum standard by itself.

Employer work sample checklist

• Identify critical job tasks and required entry-level competencies.
• Create a blueprint linking each task to scoring criteria.
• Use realistic, bounded and non-confidential materials.
• Provide clear instructions, tools and time allowances.
• Prepare answer keys and behavioural rubrics.
• Define partial-credit and error-handling rules.
• Calibrate assessors and review scoring consistency.
• Pilot for timing, usability and accessibility.
• Justify weights, cutoffs and decision rules.
• Maintain version control and monitor results.

How Talentrenza Assess can support practical selection

Talentrenza Assess provides structured assessment workflows that employers can use alongside job-related practical exercises and other selection methods. Employers should verify which exercise formats and scoring capabilities are supported before representing a work sample as automatically administered or marked by the platform.

A platform can organise assessment delivery and results, but the employer remains responsible for task relevance, scoring standards, candidate fairness and the interpretation of scores.

For related guidance, see Design Pre-Employment Assessment Tests by Job Role and Pre-Employment Assessment Pass Marks and Scoring Criteria.

Frequently asked questions

What is an example of a work sample test?

An accounts payable candidate might reconcile a fictional supplier statement, identify a ₦27,000 discrepancy and explain the required follow-up. The output can be scored against an answer key and rubric.

Are work sample tests better than interviews?

They provide different kinds of evidence. A practical task can demonstrate performance on a defined activity, while a structured interview can explore reasoning and communication. Their value depends on job relevance and implementation.

How long should a work sample assessment take?

There is no universal duration. Use the shortest exercise that provides sufficient relevant evidence for the decision, and pilot the time allowance.

Should employers pay candidates for work samples?

Policies and legal obligations vary. Employers should avoid extracting usable commercial work from applicants and consider compensation when tasks are extensive or produce business value.

Can work samples be automatically scored?

Some outputs, such as numerical calculations or objective technical answers, can be scored automatically. Complex written or behavioural responses may require trained human review or separately validated scoring procedures.

Does a realistic work sample guarantee predictive validity?

No. Employers still need evidence about coverage, scoring reliability, fairness and the intended use of results.

Conclusion

A useful work sample is a focused demonstration of important job skills, not a test of endurance or a source of unpaid business output. Start with the role, design a bounded task, define scoring rules, pilot the experience and interpret results with appropriate caution.

When employers can explain what the candidate demonstrated, how it was scored and why the evidence matters for the job, work sample tests become a more transparent part of a defensible hiring process.