Why Your Turnitin Similarity Score Changed After the Due Date — the Regeneration Mechanism

Your similarity score can change after the due date because Turnitin effectively runs the comparison twice. At submission, the report compares your paper against the database of previous submissions, journals, and internet sources — but not against the other students in your class. After the due date, Turnitin regenerates reports for the whole assignment and adds matches between students. University FAQ pages describe it as: the similarity report "is completely and automatically re-generated for ALL papers in the assignment" after the due date. That is why a score can move from 8% at submission to 25% later — not because anyone altered your paper, but because the later report includes a category of matches the first one did not.

The Portsmouth university guide puts the mechanism in one sentence:

"Turnitin reports run twice. Once upon submission. Once when the dropbox closes."

The first run answers "how much of my paper matches existing sources?" The second run answers "how much matches existing sources plus my classmates?" UCL's student FAQ states the first run directly:

"The similarity report that is immediately generated and provided to the student, before the assignment Due Date does NOT include matches to other student papers in the same assignment. Before the assignment Due Date passes, the similarity report only includes matches to existing student papers within the Turnitin database, journals and publications, and other internet sources."

So a low score at submission is not a mistake — it is an incomplete view by design. The full view arrives later.

At the due date, the second run happens for everyone. UCL's FAQ continues:

"After the Due Date has passed, the similarity report is completely and automatically re-generated for ALL papers in the assignment, thus also comparing papers between students in the same assignment. This may result in the similarity percentage that is displayed after the Due Date differing from the result displayed at the time of submission."

Turnitin's own student guide describes the same final check from the collusion angle:

"Turnitin can identify that collusion has taken place in this scenario by running a final similarity check against all submitted assignments after the due date. This ensures that every student is subject to the same level of scrutiny, regardless of when they submitted their assignments."

The key words: completely, automatically, for ALL papers. The regeneration is not targeted at you, and it is not a manual action. Everybody in the assignment gets the same final check.

The most common reason a score jumps at the due date is not copying — it is common text. Western Sydney University's FAQ names the scenario:

"Students may notice an increase in similarity after the due date if an assignment template has been used. In a situation where students have used template text such as table headers or report sections, these sections of text will only show as matches within the class group after the due date has passed."

Think about how that works: the template text on the instructions page, table headers, boilerplate sections — every student is using the same text. At submission time, that text does not match other classmates because classmates are not in the comparison. After the due date, the whole class is compared, the shared template text matches across everyone's papers, and the whole class's scores can rise together. This also happens when two or more students independently include the same quoted passage or the same source material — the classroom comparison surface grows at the deadline.

Two things stay fixed, and both are worth knowing because they are frequent sources of worry:

• Late submissions do not regenerate anyone else's report. UCL: "The assignment will NOT regenerate reports for the already-submitted papers just because a student submits late." If a classmate submits after the deadline, the reports of the on-time papers are not recalculated to include that late paper. Western Sydney says the same: reports of on-time papers "will not consider or include the late submission."
• The submission-time report was real. The 8% you saw at submission was an accurate report of one thing — matches against sources outside your class. It did not falsely claim there were no matches; it reported the comparison that existed at that moment.

Reading this mechanism changes how you should react to a score change:

1. Do not assume tampering. The regeneration is automatic and class-wide. If your score moved, the mechanism did — nobody edited your paper.
2. Know the two numbers mean different things. Submission score = matches against the external database. Post-due-date score = external matches plus classmate matches. They are both real; they are different questions.
3. Expect the jump if you share visible text. If your class used a template, or you and others quoted the same passage, a rise at the due date is the mechanism, not a verdict.
4. If you believe something is wrong, ask your instructor. The report is their investigative tool; they can open your report and explain what changed and why — and they are the person who interprets it.

A similarity score changing after the due date is the documented behavior of the tool, not a mystery and not tampering. The report runs twice: at submission (external sources only, no classmates) and at the deadline (regenerated for all papers, classmates included). Common shared text — templates, headings, quoted passages — is exactly what makes the whole class's scores rise together. Late submissions do not trigger a new run for on-time papers. The two numbers are two different views of your paper, both accurate for what they compared. If you see your score move after the deadline, you now know what happened — and if it still worries you, ask your instructor to walk you through your report.