Evaluator guide
How automatic grading works
EvaluatIA compares responses with each question’s saved definition. Compatible objective formats receive automatic scores; responses requiring judgment remain pending until a valid review exists.
- Automatic
- Compatible objective formats
- Pending
- Responses requiring judgment
- Final grade
- Only when no reviews remain pending
What can be graded automatically
Deterministic grading supports single choice, multiple choice, true/false, fill-in-the-blank, matching and ordering, provided a valid correct-answer definition is retained.
Some text or math responses can also be graded automatically with a compatible mode and expected answers. Inconclusive comparisons go to review.
How each answer is compared
- Single choice and true/false: compare the selected option to the correct option.
- Multiple choice: all correct options must match with no additional options.
- Fill-in-the-blank: apply the configured text normalization rules.
- Matching: each item must retain the expected correspondence.
- Ordering: the entire sequence must match the saved order.
What remains pending review
Short, open, media analysis and programming questions require manual review when an answer lacks a valid final grade. Short answers are auto-graded only with an explicit, verifiable mathematical or numerical mode.
AI can suggest grading for short, open and media analysis responses. Programming is outside that assistance, and the evaluator retains the final decision.
While a response is pending or a score is invalid, the attempt remains Pending review and is not presented as a complete final grade.
How to check results
Open Results
Find the assessment and check for pending responses.
Open the attempt
View answers, automatic scores and each question’s maximum.
Grade pending responses
Save a valid score and the appropriate feedback.
Finish review
When no responses remain pending, EvaluatIA can close the attempt’s grading cycle.
Direct answers
Frequently asked questions
Does a blank answer remain pending?
Automatic formats usually score zero. A manual response without reviewable content does not alone block review; the state also depends on valid scores for every expected question.
Does AI automatically grade all open responses?
No. AI can suggest scores for short, open and media analysis answers, but the evaluator accepts or changes them. Programming is not part of AI grading.
Continue in EvaluatIA
Open the feature for your activity. Evaluator features require an account and completion of initial setup.