One of the most common anxieties for authors submitting to IEEE conferences or journals is the Similarity Index (often called the "查重率" or plagiarism score).

If your manuscript has hit a 30% similarity score, you are in a high-risk zone. While IEEE does not have a single "magic number" that applies to every society, a score above 30% will almost certainly trigger a manual investigation and significantly increases the chance of "Desk Rejection" (rejection without review).

Here is the breakdown of how IEEE editors view your score.

1. The General Thresholds (The "Traffic Light" System)

Most IEEE publications use CrossCheck (powered by iThenticate). While specific tolerances vary by journal/conference, the general industry standards are:

• Safe Zone (< 15% - 20%):
• If your total similarity is below 20%, you are generally safe, provided there are no large blocks of copied text from a single source.
•  Warning Zone (20% - 30%):
• This range usually triggers a Manual Review. The editor will look at what is being flagged.
• If the matches are mostly in the "Methodology" or "References" (common technical phrases), they may ask you to revise it.
• If the matches are in the "Abstract" or "Conclusion," it looks suspicious.
•  Danger Zone (> 30%):
• High Risk of Rejection. A score this high usually indicates that entire paragraphs have been copied.
• Immediate Action: Do not submit a paper with >30% similarity. You must paraphrase and rewrite to bring it down.

2. "Total" vs. "Single Source" Similarity

The Total Similarity (30%) is not the only number that matters. The Single Source Similarity is often more critical.

• The Rule of 5%: Even if your total score is only 15%, if 5% or more comes from one single paper, that is a major red flag. It suggests you are heavily relying on (or copying) one specific work.
• IEEE Standard: Most IEEE editors want to see no more than 1% - 3% match from any single source.

3. The "Self-Plagiarism" Exception

There is one scenario where >30% might be acceptable: Extending a Conference Paper into a Journal Paper.

• If you are submitting an expanded version of your own previously published IEEE conference paper to an IEEE Transaction/Journal, a higher similarity score (e.g., 30%) is often expected.
• Requirement: You must cite your original conference paper and explicitly state that this is an "Extended Version" (usually requiring 30-40% new content).

4. How to Fix a High Score

If your draft is sitting at 30%, follow these steps before submitting:

1. Exclude References: Ensure the check excludes the bibliography (references).
2. Paraphrase Methodology: This is where most "accidental" plagiarism happens. Rewrite standard procedures in your own voice.
3. Synthesize, Don't Quote: In engineering, direct quotes are rare. Read the source, close it, and write the idea from memory to ensure originality.

Summary

Will >30% be rejected? Likely, yes. Or at the very least, it will be sent back to you for major revisions immediately. To ensure a smooth review process, aim for a Total Similarity < 20% and a Single Source Similarity < 3%.