Help:Earwig's copyvio detector
Earwig’s Copyvio Detector is an external tool designed to help contributors identify potential issues of non-original content in articles. It detects similarity between the text of a target page, such as a Wikinews article and content found online. This can include:
- Pages linked from the article (including the Sources section)
- Pages returned by a search query based on the article’s content
What the tool does
Earwig’s tool provides a percentage similarity score. This number indicates how much of the article matches other content online, but:
- It does not measure the percentage of copyright violation.
- It can detect quotations, even if properly attributed, and report them as similar.
- It may flag source titles (e.g., in the Sources section) as part of the similarity, especially if they resemble the article’s title (as they should).
- It has no direct correlation to or foundation in Wikinews policies and/or guidelines
What the tool does not do
Earwig’s tool does not determine whether matched content is copyrighted. Contributors must manually check if a source is:
- Public domain
- Freely licensed
- Claimed under fair use
- Or copyrighted and restricted
Even if content is free to use, it can still be plagiarized if copied without attribution or presented as original. Plagiarism concerns authorship; copyright concerns legal rights. Both matter on Wikinews.
A note on "scuffing"
Superficially rewording source material, often called "scuffing," is not enough to avoid plagiarism. Changing a few words or rearranging sentences without producing original writing still misrepresents authorship and violates Wikinews standards.
In Earwig’s report, scuffed content may show up as:
- A low to moderate similarity percentage, even when much of the paragraph closely mirrors a source.
- Multiple short red highlights scattered through a paragraph, rather than one continuous block. These may be common phrases or key terms left unchanged.
- A source that shows many “Compare” matches, but no single one appears severe—yet taken together, they suggest the structure, flow, and word choice are not original.
Always aim to write in your own words, based on a full understanding of the source, not by mimicking its form.
Human judgment is required
Earwig’s tool is only a guide. It is not definitive. Reviewers and authors must use editorial judgment to determine whether content is:
Plagiarism - using someone else’s work and presenting it as one’s own without proper attribution.
Copyright violation (copyvio) — reusing copyrighted material without permission or proper legal basis.
Even when the similarity score is high, properly quoted and sourced material is usually not plagiarism or copyvio.
Understanding Earwig’s output
Earwig presents its results in a way designed to help users compare the article’s content with potential sources in a side-by-side manner.
A permalink example of a result is available here. It can be used as a reference for the following description:
Checked sources panel
Below the panel reporting the percent similarity is the “Checked sources” section. This lists the web pages Earwig examined for similarity to the target article. For each source, it displays:
- The percentage similarity to the article
- A “Compare” link that opens a direct side-by-side comparison
Depending on the settings used, these sources are typically drawn from:
- Pages cited in the article’s Sources section
- Search results found by querying portions of the article text
Article vs Source pane
Beneath the “Checked sources” panel are two side-by-side panes:
The left pane, labeled Article, shows excerpts from the Wikinews article
The right pane, labeled Source, shows matched content from external web sources
Phrases highlighted in red in both panes are identical strings of text. These show verbatim matches between the article and a source.
The right (Source) panel may display content from multiple sources. As you scroll down, each new source is clearly introduced with its title, allowing you to track which parts of the article match which source.
Action required when seeing plagiarism and/or copyright violations
Contributors are strongly encouraged to check the article before submitting for review by clicking the 'Earwig' link in the 'developing' template at the top of the article. This link will automatically start an evaluation by Earwig and report the results in the same page.
Once Earwig flags similar phrases, each should be reviewed for potential plagiarism or copyvio. Phrases that are formal titles, widely used expressions, or routine factual language are typically not violations.
To deal with plagiarism or copyvio:
- The offending phrase(s) needs to be written in your own words or removed. If you don't have time to do this, leave a note on the talk/collaboration page of the article.
- If you are not the original author, notify them.
- If the article is already edit protected, request an admin to address the text using {{Editprotected}} in the article talk space.
- Note for reviewers and administrators: If plagiarism or copyvio is removed from a published article, a {{Correction}} notice must also be added to the article.
If the entire article is a copyright violation or plagiarism, then
- Blank the article and tag it using {{Copyvio}}.
- Notify the author at their personal talk page using {{Copyvio notice}}.
Bottom line
Earwig is a helpful tool for detecting textual similarity, but it doesn’t judge plagiarism or copyright violation. High similarity doesn’t always mean copyvio and low similarity doesn’t guarantee originality.
Plagiarism is about misrepresenting authorship.
Copyvio is about reusing protected content without permission.
Use Earwig as a guide, but always apply human judgment. And write articles using original language.