Understand reports

Overall similarity

The overall percentage summarizes overlap between the submitted text and available configured sources. It is not a universal threshold for misconduct. Appropriate interpretation depends on task type, permitted collaboration, citation practice, templates, discipline, and source availability.

Source details

ScholarScan can compare against:

• Prior indexed submissions in the same institution tenant.
• Institutional reference documents.
• Lecturer-local reference documents.
• Online sources when a provider is configured.

The report lists matched passages and source-level measures. Inspect the language around every significant match. Similarity caused by a reference list, quotation, required phrase, or common technical language may be legitimate.

AI writing indicators

The AI section reports status, risk, score, confidence, reasons, and weighted language signals. Moodle-style help popovers explain each calculation and provide a research reference.

Signals currently include sentence uniformity, moving-average lexical diversity, repeated word sequences, repeated sentence openings, transition density, and hedging language. None uniquely identifies AI authorship.

Danger

Do not make a disciplinary finding from an AI score. Discuss the work with the student and consider drafts, version history, oral explanation, source use, language background, accessibility needs, and assignment constraints.

Confidence and insufficient text

Confidence reflects whether there is enough text and sentence structure to calculate the signals reliably. Insufficient text is a limitation of the analysis, not evidence for or against AI use.

Grading suggestion

When Suggest grade from lecturer files is enabled, ScholarScan can provide a percentage based on the closest lecturer reference match. This is a suggestion only. It does not update Moodle grades and should not replace rubric-based assessment.

A defensible review record

Document:

• Which report and submission version was reviewed.
• Which source passages were material.
• Which exclusions and source collections were enabled.
• Any discussion with the student.
• Evidence beyond automated indicators.
• The policy and academic judgement applied.