Documentation / QueryTek Review glossary / Scorecard
A scorecard is the completed record of one review: each of the rubric's dimensions with the level assigned to it, the evidence supporting that level, and the resulting outcome. The rubric is the instrument; the scorecard is the reading it produced, and the artefact that outlives the reviewer's memory.
A scorecard is worth more than the score on it, because the score alone cannot be checked. Three elements make it reviewable.
Per-dimension results rather than a single figure. An overall score of 3 out of 5 conceals whether the material is uniformly adequate or excellent in substance and unusable in clarity — which are different problems with different remedies.
Evidence links. Each dimension's level should connect to the annotations supporting it. This is what allows someone to disagree specifically rather than in general, and what makes the outcome defensible months later when nobody remembers the document.
The decision, stated separately. Dimension levels are observations; the outcome is a judgement. Whether material with a weak clarity score is returned or accepted with notes depends on the engagement's thresholds, and recording that decision as its own element keeps the observation distinct from the consequence.
Scorecards become more useful in aggregate. Across an engagement they show which dimensions consistently fall short, which is more actionable than knowing the average — a program where clarity always scores lowest has a systematic authoring issue, not a review issue.
Aggregation has a limit worth respecting. Scorecards are comparable when they share a rubric; across rubrics the numbers use the same scale to mean different things. Reporting that averages across engagements with different profiles produces a figure with no interpretation.
Used carelessly, scorecards also invite reviewer comparison. Reviewers handling harder material score lower, so raw score differences measure the material as much as the reviewer.
A completed review needs a durable artefact. A verbal or informal outcome cannot be revisited, aggregated, or defended, and it disappears when the reviewer moves on.
Scorecards also turn review into something a program can learn from. Recurring low dimensions point at where material is going wrong upstream, which is usually cheaper to fix than reviewing harder.
QueryTek Review records outcomes as per-dimension scorecards linked to supporting annotations, so a decision can be traced to its evidence. Scorecard contents belong to the engagement and are not published.
