In this article, we explain how to interpret Turnitin similarity scores from the similarity report. This article has the following structure:
Opening a similarity report
To find the similarity report:
- Open the Assignments tool via Course Tools. Here you will see all assignments (formative and summative), all student submissions, and grading information. Click on the assignment for which you want to view the similarity report.
- The Turnitin Similarity column displays all generated reports.
- To open a similarity report, click on the percentage in the Turnitin Similarity column.
The similarity report will open in a new window.
How does the similarity report work?
The assignments submitted by students are compared with everything in the Turnitin database. This database consists of internet content, previously submitted assignments in Turnitin (including those submitted in Blackboard in previous years), academic publications, and more. If similarities are found with passages from the database, these will be highlighted in the similarity report. Similarity reports provide an overview of the sections of text in the submitted assignment that match text from the database. See also this explanation on the Turnitin website.
A high score does not necessarily mean that plagiarism has occurred. Similarities can also result from the correct use of quotations. As an instructor, you determine whether plagiarism has actually occurred.
The colors in the Turnitin Similarity column indicate the level of similarity. This allows you to quickly see whether there are many or few matches found.
- Blue: No matching text found
- Green: One word to 24% matching text
- Yellow: 25-49% matching text
- Orange: 50-74% matching text
- Red: 75-100% matching text

Similarity Report: An overview
After clicking on the percentage in the Turnitin Similarity column, you can view the details in the similarity report. In the report, the text with matches is highlighted. The highlighted passages have a number that refers to the source(s) where the match was found. By clicking on the number, you can view the source publication and assess whether the match constitutes plagiarism.

Viewing the detailed report
This article assumes the 'New view.' If the layout does not match, click on Switch to the new view.

- Overall similarity: Shows the number of text matches found as a percentage.
In this example, 87% of the submitted product matches comparison sources from Turnitin.
From the overall similarity, it cannot be automatically concluded that the submitted work contains plagiarism. With a high similarity index, the text matches must be assessed to determine whether they are indeed undesirable.
- This list shows the sources with which the work has similarities.
- If you want to see all sources in one list, you can turn on Show overlapping sources.
Click on the individual sources to compare the text in the source with the text in the product with the corresponding number:

- View other sources: It is possible that a section of text shows similarities with multiple sources.
The source with the most matches is always displayed first. When multiple sources have the same number of matching words, they will be displayed in the following order:
- Internet sources
- Publications
- Submitted works (student papers)
Setting filters and excluding matches
To refine the similarity scores, you can use filters and/or exclude matches.
More information about this can be found on the Turnitin website.

If you want to exclude an individual source, you can do so in the following way:
- Go to the source you want to exclude.
- Click on the 'Exclude Source' icon.

Integrity flags
In addition to text matches, Turnitin also indicates when text contains suspicious symbols and formatting using Integrity Flags..

There are two types of integrity flags:
Hidden text: This is a way in which similarity matching can be bypassed, and plagiarism may not be detected.
Similar characters: Sometimes characters look so similar that a student can use these characters to bypass the plagiarism check. Turnitin then reports them as similar characters.