Blog / Work Samples: Separate the Deliverable From the Behaviour Observed While Producing It
Work Samples: Separate the Deliverable From the Behaviour Observed While Producing It
A polished task output and a recorded process are different evidence from the same exercise.
October 2, 2026 · 6 min read · By Cepingku Editorial Team
Ready to start?
Start the Management Development →A work sample is not simply any piece of completed work labelled as proof of ability. Its value as an exercise depends on asking people to handle tasks and conditions related to the work of interest, with the evidence to be scored defined in advance. The U.S. Office of Personnel Management explains that work samples and simulations can be rated from task outputs or from behaviour seen by trained observers. For hiring, it also notes that this approach fits skills needed at entry and may be unsuitable when those skills will be taught after selection. That describes how the method can collect evidence; it does not establish predictive validity for this site’s assessment or for a generic practice task.
A finished document tells you what was delivered. A process observation tells you what was visible while the person worked under the exercise conditions. The two sources can complement each other, but neither automatically settles the other. A clean output does not prove that a person manages a live team well, and an unfinished output does not identify why it was unfinished without looking at the task conditions and the available record.
Start with one fictional, comparable exercise
Consider this fictional development exercise, not a validated scale or a hiring screen. Each participant receives the same one-page project brief. With 20 minutes and no outside assistance, they draft a priority order and a handoff note. The brief already names two dependencies, three pending tasks, and one conflict between deadlines. Before anyone begins, the exercise author defines the material, time limit, permitted help, required output, and rating anchors.
| Condition fixed in advance | Why it matters |
|---|---|
| Same one-page brief and known information | A participant with more background should not be ranked directly against one without it. |
| Twenty minutes and no outside help | Time pressure and available resources are held comparable. |
| Priority order plus handoff note | The requested deliverable is defined rather than left open-ended. |
| Rating anchors | Observers can point to specific records instead of relying on a general impression. |
These controls do not turn the exercise into a miniature of every real job. They set the boundary of this comparison. A longer time limit, different materials, or permission to consult a colleague creates a different exercise. Its output should not be placed on the same ranking without a stated basis for doing so.
The output and the process answer different questions
In this fictional exercise, output anchors could ask whether the written priority order covers the two known dependencies, whether the handoff identifies an owner and deadline, and whether it leaves a traceable note about the conflict. Those criteria inspect the submitted text. Behaviour anchors could ask whether the participant identifies an important unknown in the brief, explains a trade-off when priorities conflict, or marks an unresolved question for confirmation. Those criteria describe acts and statements that a trained observer can record during the allotted exercise.
| Evidence type | Record that can be checked in this exercise |
|---|---|
| Task output | Does the handoff cover both known dependencies and identify an owner and deadline? |
| Behaviour observation | Did the participant flag an unknown and explain a trade-off between conflicting priorities? |
Not asking a question does not necessarily mean that a person cannot ask questions. The rules may provide no one to ask, or may forbid external help. The accurate observation is narrower: no unknown was raised within these materials and this time limit. Likewise, a polished handoff note does not establish that its author will coordinate, delegate, or respond to surprises in a real team. The record should name the opportunity the exercise supplied, then stop at what it supports.
Reconcile observations; do not average impressions into certification
If two observers take part, each can first write records against the same anchors: for example, ‘included dependency A and B’ or ‘listed the deadline conflict as needing confirmation.’ They can then compare their records and revisit the material. A disagreement may expose an unclear anchor, an incomplete record, or an exercise that never gave an adequate chance to show the behaviour. The first repair is to improve the task, opportunity, materials, time limit, or anchor for a future exercise.
Averaging two broad impressions does not transform a practice exercise into a certified management score. The fictional choices in this article are editorial teaching devices, not a validated selection instrument. This site’s Management Development assessment can prompt reflection on planning, coordinating, and developing people; it does not use these exercise anchors and should not be interpreted as an OPM work sample, a hiring conclusion, or a performance credential.
Want to see where you land?
Take the Management Development and get your results in minutes.
Take the Management Development →