Plagiarism in Research Papers: Types, Detection & How to Stay Under the Threshold
Most authors think of plagiarism as one thing: copying someone else's sentences. In practice, similarity checking tools and editorial review catch several distinct patterns, some of which authors commit accidentally while genuinely trying to do the right thing.
Types of Plagiarism Authors Don't Always Recognise
- Direct copying — reproducing text word-for-word without quotation marks or citation
- Poor paraphrasing — swapping a few synonyms while keeping the original sentence structure intact
- Self-plagiarism — reusing your own previously published text substantially without disclosure
- Mosaic plagiarism — piecing together phrases from multiple sources without adequate attribution, even if no single sentence is copied wholesale
- Uncited data or figures — reusing a table, chart or dataset from another source without crediting it
How Similarity Checking Actually Works
Tools like Turnitin and iThenticate compare your manuscript against a vast database of published papers, web content, and previously submitted work, then generate a similarity percentage showing how much text matches existing sources. This isn't a plagiarism verdict by itself — it's a starting point for a human reviewer to examine flagged sections.
Why 15% Is a Common Threshold, Not a Magic Number
IJRTI, like many journals, requires a similarity score below 15% before a manuscript proceeds to peer review. That number isn't arbitrary cruelty — it accounts for the reality that some matched text is unavoidable and legitimate: standard methodology phrasing, properly quoted material, common technical terms, and reference list entries all register as "matches" even in entirely original work.
When a High Score Isn't Actually Plagiarism
| Flagged As | Often Actually Is |
|---|---|
| High match on references section | Standard citation formatting matching other papers' reference lists |
| Match on methodology description | Widely-used standard procedure phrasing (e.g. common statistical test descriptions) |
| Match on quoted text | Properly quoted and cited material, correctly flagged as a match but not an ethics issue |
| Match against your own prior paper | Requires disclosure as related/prior work, not automatically self-plagiarism if properly cited |
This is exactly why a human reviews the report rather than an automated cutoff alone deciding rejection.
How to Legitimately Reduce Your Similarity Score
- Paraphrase genuinely — restructure the sentence and use your own framing, not just synonym swaps, then cite the source
- Use quotation marks and a citation for any text you do keep verbatim
- Disclose and cite your own prior work if you're building on it, rather than silently reusing text
- Write your literature review in your own analytical voice, summarising and comparing sources rather than stitching quotes together
What Happens If Plagiarism Is Confirmed
A manuscript with a confirmed plagiarism issue before publication is returned for revision or rejected, depending on severity. If confirmed after publication, it's treated as research misconduct under IJRTI's COPE Ethics Policy, which can lead to retraction — full details of that flowchart are documented there.