Turnitin uses a much larger database and a more advanced detection technique than Ouriginal. As a result, the system recognizes more similarities, including short text fragments or paraphrased sentences, which can lead to higher scores.

The detection quality is better, but a higher score does not automatically mean plagiarism. Turnitin only signals overlap — it is up to the instructor to judge whether it concerns correct citations, standard formulations, or unauthorized reuse.

Brightspace

Go to More options → Submission settings and select Do not store submitted papers. This will keep the document confidential and prevent it from being used or compared in future submissions.

Turnitin Stand-alone

At the bottom of the Customize Your Search page, go to Submit papers to: and select No Repository.

If you want to have a document checked without it being added to the Turnitin database, you can set this when creating the assignment in Brightspace.

Go to More options → Submission settings and select Do not store the submitted papers. This ensures the document remains confidential and is not used or compared in future submissions.

As an instructor, you can use filters to refine the score in Turnitin.
When creating an assignment in Brightspace or submit via the Stand-alone version , go to More options → Similarity Report and select what you want to exclude: Bibliography, Quotes, and optionally Small Matches (short similarities).

After submission, you can open the Similarity Report and use the filter icon (funnel) to adjust the same settings and immediately see the effect on the score.

Turnitin compares the submitted document with a large collection of public sources, such as websites, academic articles, and student papers from other institutions. Only the overlapping text is shown in the report — the original source is not accessible to the instructor but is used solely to indicate overlap.

Turnitin supports more than 30 languages, including Dutch and English. Detection works best for languages with many online and academic resources. For less common languages, the similarity score may be lower because there is less material to compare against.

Turnitin also recognizes text in programming languages, such as Python or Java, but does not assess functionality or structure — only literal similarities in code fragments.

Yes, by default, references and bibliographies are included in the similarity score. Instructors can change this: when creating the assignment in Brightspace or submit via the Stand-alone version go to More options → Similarity Report and check Exclude Bibliography.
This way, citations and references are not counted in the score, resulting in a more realistic report.

Turnitin recognizes standard phrases and short sentences that frequently appear in texts but usually does not count them as plagiarism. Only longer or unique text fragments are marked as potential matches. The instructor can also choose to exclude small matches via More options → Similarity Report → Exclude Small Matches.

Turnitin can only perform plagiarism detection if a document contains enough text. Files with fewer than 20 words are not scanned because there is too little content for a reliable comparison. The more text the document contains, the more accurate the similarity analysis will be.

When a student submits a document to an assignment where Store submitted papers is enabled, the file is added to the Turnitin database after processing. This usually happens within a few minutes. From that point, the document can be used as a comparison source for new submissions.

Yes, Turnitin can scan preview content from paywalled websites, but only the part that is publicly visible — for example, an abstract or the first paragraph of an article. The full paid content behind the paywall is not indexed or used for comparison.

At the moment, library contracts are separate from the sources that Turnitin can check for similarities. However, Turnitin has significantly more licensing and publisher agreements than Ouriginal — including contracts with Elsevier — which means that in practice, it can identify a much larger number of academic publications and articles.

No, Turnitin does not compare the documents within a zip file to each other. Each file is analyzed separately and compared only against Turnitin’s external databases. If you want student work to be compared internally, the files must be submitted individually.

Documents are only compared with each other if they are stored in the Turnitin database. This happens when the Store submitted papers setting is enabled when creating the assignment.
If Do not store the submitted papers is selected, the documents are not added to the database and therefore not compared with each other.

There needs to be some time between two submissions if the documents are to be compared with each other. When two documents are submitted almost simultaneously, they may not yet have been processed into the database.
It is advisable to wait until a report has been generated for the first submission before submitting the next one via Quick Submit.

In submissions via Brightspace, Turnitin determines automatically when documents are added to the database. This usually happens shortly after submission, but the exact processing time can vary.
Instructors and students have no direct control over this: if two documents are submitted almost at the same time, the comparison between them may occur during a later scan.
In later submissions (for example, resubmissions or new versions), earlier submissions are usually recognized and compared automatically once the first version has been processed in the database.

When a document is deleted from the Turnitin database, it is completely removed from the comparison archives. This means it can no longer be used as a source in future plagiarism checks.
Existing reports in which the document already appeared as a match remain visible, but new submissions will no longer be compared against it. The text itself is then permanently no longer searchable within Turnitin.

Permanent deletion can only be requested through the key user and is performed by the vendor.