> ## Documentation Index
> Fetch the complete documentation index at: https://docs.exterview.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Scoring Quality

> Keep Exterview scoring consistent across reviewers with structured rubrics, calibration practices, and clear reviewer expectations for every evaluation.

## Purpose

Consistent scoring is what makes an evaluation defensible. These practices keep scores comparable across candidates and reviewers.

## Scoring hygiene

<CardGroup cols={2}>
  <Card title="Shared rubrics, clear levels" icon="list-check">
    Define what each score level means so reviewers apply it the same way.
  </Card>

  <Card title="Calibrate reviewers" icon="scale-balanced">
    The same inputs should produce the same score, regardless of who reviews.
  </Card>

  <Card title="Escalate, don't round" icon="triangle-exclamation">
    Send borderline candidates to human review instead of forcing a score.
  </Card>

  <Card title="Monitor outcomes over time" icon="chart-line">
    Check that scores still predict what actually matters after hire.
  </Card>
</CardGroup>

## The platform features that support this

<CardGroup cols={3}>
  <Card title="Scoring Rubrics" icon="list-check" href="/configure/operations">
    Structured dimensions, configured once and applied consistently.
  </Card>

  <Card title="Calibration" icon="scale-balanced" href="/monitor/calibration">
    Versioned scoring so results stay reproducible over time.
  </Card>

  <Card title="Auditability" icon="clipboard-list" href="/trust/auditability">
    Every score traces to the rubric version and the reviewer behind it.
  </Card>
</CardGroup>

## Scoring quality checklist

| # | Confirm                                         |
| - | ----------------------------------------------- |
| 1 | Criteria are defined for each rubric dimension. |
| 2 | Reviewers understand the rubric.                |
| 3 | Reports are reviewed consistently.              |
| 4 | Exceptions are escalated, not absorbed.         |
| 5 | Outcomes are monitored over time.               |

## FAQs

<AccordionGroup>
  <Accordion title="How do we keep two reviewers consistent?">
    Shared rubrics with clear level definitions, plus calibration so the same inputs score the same way.
  </Accordion>

  <Accordion title="What should we do with borderline candidates?">
    Escalate them for human review rather than forcing a score. Consistency comes
    from process, not rounding.
  </Accordion>

  <Accordion title="Does Exterview expose the exact weights?">
    No. The methodology is explained; the specific weights and thresholds stay in your signed configuration.
  </Accordion>
</AccordionGroup>

## Related

<CardGroup cols={2}>
  <Card title="Evaluation Design" icon="bullseye" href="/best-practices/evaluation-design">
    Design the evaluation before you tune how it's scored.
  </Card>

  <Card title="Analytics" icon="chart-line" href="/monitor/analytics">
    Track scoring and review activity across your whole pipeline.
  </Card>
</CardGroup>