How assessment scores are calculated
An assessment score is built from the answers to its questions. Understanding how each question type contributes is the difference between an assessment that measures something and one that produces a number everybody ignores.
Where: Performance › Assessments You need: performance rights to view assessments
The scoring model
Each question contributes a score based on the option chosen. Options are ordered, and the score reflects where the chosen answer sits in that order — which is why the order of your answer options matters as much as their wording.
Write options from most positive to least positive, consistently, across every question. An assessment where some questions run best-to-worst and others run worst-to-best produces scores that are arithmetically correct and meaningless.
Multiple-choice questions always score maximum
⚠️ A multi-select question awards the maximum score regardless of what is selected. It is not scored against the options chosen. In KAMI's own words, in the scoring code: "No scoring for multiple choice, just always assume max."
This is the single most important thing on this page. If you include multi-select questions in a scored assessment, every employee receives full marks for them, and those questions cannot distinguish anybody from anybody else. Worse, they inflate every score by the same amount, so the assessment looks like it is working.
Use multi-select for information you want to gather, not for anything you intend to score. If a question should affect the score, make it single-answer with ordered options.
Questions with no options
A question with no answer options contributes nothing to the score. A question whose options do not extend beyond the first contributes the maximum.
Both are worth knowing when an assessment's scores look higher or flatter than expected — the cause is often the shape of the questions rather than the answers.
Ratings
Reviewer ratings run from 1 to 5. That scale is what feeds the wider performance reporting, including the talent grid — see Performance scores and the talent grid.
What Happens Next
- The score is recalculated from the current answers, so it reflects the assessment as it stands.
- Scores feed team reporting and the talent grid.
- Changing a question's options after an assessment has been answered changes how it scores — avoid mid-cycle edits to a live cycle.
Tips
- Audit your question set for multi-select before running a cycle. This is the check that most improves the quality of your scores, and it takes minutes.
- Order every question's options the same way. Most positive first, consistently. Mixed direction is the commonest cause of scores nobody trusts.
- Do not edit questions mid-cycle. Employees answered a different question from the one you are now scoring.
- Sanity-check the distribution. If nearly everyone scores near the top, look for multi-select questions and single-option questions before concluding your team is uniformly excellent.
- Keep scored and informational questions separate in your own design notes, so the next person building a cycle does not mix them.
Troubleshooting / FAQ
Q: Everyone scored very highly. Check for multi-select questions, which always award maximum, and for questions with only one option.
Q: A question does not seem to affect the score. A question with no options contributes nothing.
Q: Two employees with different answers got the same score. If the differing answers were on multi-select questions, that is expected — those do not discriminate.
Q: Can we change the scoring weights? Scoring comes from the options and their order. Design the options to express the weighting you want.
Q: We edited a question mid-cycle. What happened to the scores? They are computed from current data, so they moved. Avoid editing a live cycle.