How the marking works
A board does not mark your story. It marks the evidence in your story against a list of indicators for the sub-dimension being tested, then places the answer in a band. The review here does the same thing, in the same order, and shows you each step.
1. Indicators, with your words quoted back
Every Civil Service grade has a published framework. Under the Capability Framework: four dimensions, eight sub-dimensions from Administrative Officer up, and for each sub-dimension a list of indicators. Under the older Competency Model: six competencies per grade, each with its own indicators. Those indicators are the behaviours a board is listening for, and we hold them as published. For each one the review says clear, partial or not evidenced, and for anything credited it quotes the exact sentence from your answer that earned it. No quote, no credit. That rule is what keeps the marking honest.
A two-minute answer normally evidences two to four indicators well. The review does not expect you to hit all of them, and neither does a board. Four clear indicators in a clean answer sits around 31, Pass (Good), which is where most successful candidates land. Very Good needs evidence well past that, and the estimate never shows 50.
With full access you can also answer the follow-up the board would ask, up to three in a row on one example. Each reply is marked together with everything you said before, the way a board hears it, and you see the mark move.
Not every question is a STAR question. Boards also ask about your skills, values and motivation, put a situation to you and ask what you would do, or test what you know about the organisation. The review switches its checklist to match: for a what-would-you-do question it looks for the order of actions, the reasoning, who you would involve and what you would check; for a knowledge question it looks for purpose, implications and your own view, and tells you to verify any figure against the organisation’s own documents.
Two rules boards are trained on are built in. Negative evidence counts: blaming colleagues, no ownership of a mistake, a real customer named. And a fluent answer with few checkable specifics is capped at adequate, because boards are warned that articulate candidates get over-rated on first impression.
2. The shape of the answer
- Scene-setting share. How much of the answer is Situation and Task. Under a quarter is ideal. Over half and the mark drops, however good the story.
- “I” share. Of the sentences describing action, how many have you as the actor. “We agreed” and “the team did” score nothing for you.
- A result the board can check. A number, a date, a proportion, a change someone noticed. “It went well” does not count.
- Learning. One sentence on what you took from it.
- Answers the question asked. A strong workload story given to an ownership question is still the wrong answer.
3. Level fit
The same example can be a pass at Clerical Officer and a fail at Higher Executive Officer. The review judges the example against a description of what the grade does: organising your own work at CO, resolving complex queries and guiding others at EO, leading people and making decisions with wider impact at HEO. Managing a task is not leading people, and boards mark the difference.
4. The mark
The checklist above is turned into a number by fixed rules, the same rules every time, so the same answer always gets the same mark and you can see exactly what moved it. One clear indicator lifts an answer out of Fail. Scene-setting over half, no “I”, no checkable result and a below-grade example each pull it down. For Civil Service grades the number lands on the published 1 to 50 scale:
Bands as published in the 2026 publicjobs candidate information booklets. HSE and council competitions use different scales; the review will use theirs when those grades go live.
5. Keep, fix, and the next question
Two things worth keeping, pointing at real sentences in your answer. Up to three fixes, written either as a line you could say or as the question you still need to answer. We never invent facts for you: if the result is missing, the fix asks you for it, it does not make one up. Then the follow-up a board would most likely ask, aimed at your weakest point. Answer that one too.
What it is not
It is not a prediction. Real boards are three people with a long day and their own reading of the indicators, and they mark the whole interview. Use the mark to find the weak answers and fix them before someone else finds them.