Assignments and Revision
Focused calculation
Question Accuracy Rate Calculator
Separate correct, incorrect, and skipped questions to report accuracy among attempts and coverage across the full set.
Enter the question dispositions
Name the denominator with the rate
Accuracy among attempts can look high when many questions were skipped. Overall correct rate keeps the full set visible.
Neither measure should be called simply accuracy without its denominator.
Before finalizing the plan, investigate whether pacing contributed to omissions.
Arithmetic behind the result
Basis: Skipped questions are excluded from attempt accuracy but remain in the overall correct and attempt rates.
Separating accuracy from coverage
Of 75 questions, 52 are correct, 17 wrong, and six skipped. Attempt accuracy is 75.4%, while the overall correct rate is 69.3% and the attempt rate is 92%.
Review mistakes and omissions separately
Wrong answers suggest attempted reasoning that needs correction; skipped questions may signal time, coverage, or uncertainty.
Those dispositions can require different revision responses.
When conditions change, return to a tool that can compare total performance across attempts.
Reconcile every question
- Reconcile every question to one status.
- Calculate attempt accuracy.
- Compare it with overall coverage.
- Classify errors by topic or cause.
Clarifying the inputs
Why exclude skipped questions from attempt accuracy?
Because no scored attempt was made; the overall rate still reflects them.
Can partial credit be included?
Not in these three dispositions. Use points or a partial-credit calculation.
Which rate belongs in a report?
State the measure that answers the question and publish its denominator.
Maintain the question-level audit
Save question identifiers or topic labels behind correct, wrong, and skipped totals. This allows later revision to target knowledge gaps separately from pacing omissions.
If partial-credit responses exist, keep them in a points analysis rather than forcing them into correct or wrong. Disposition rules should remain consistent across attempts.
If the status counts do not sum to the full set, label the remainder—for example, unscored, excluded, or partially credited—before publishing any overall rate. The validation permits a remainder because some workflows include such cases, but the record should explain it.
For timed practice, compare skipped-question topics with the clock position when each omission occurred. This helps distinguish missing knowledge from a late-section pacing problem.
Accuracy can improve while coverage falls, so retain both rates across every attempt.
If a question is excluded after test review, remove it from the total as well as from every disposition count. Leaving it in the denominator silently depresses the overall correct and attempt rates.
When comparing attempts, apply the same rules for multi-part, corrected, and invalid questions so status changes reflect performance rather than classification drift.