Dear Fikrat, 

The acceptable percentage could vary depending on sources of the similar texts.

 According to springer guideline (see attached), editors should  Interprete reports and Read Percentages carefully. A high similarity percentage should be further investigated for:

Text from Many Individual Sources

How many individual sources are linked to this paper? The best way to check this is to use the Content  Tracking mode in iThenticate. A paper with single sources where each source shows similarity of only 1-5%, is generally showing no sign of potential plagiarism, provided that the majority of the paper is not made up of borrowed strings of material.

High Percentage from a Single Source

Some papers with a low overall similarity percentage can have a high percentage from a single source,  for example, the overall percentage is 15%, but 12% is coming from one single source. Generally, the  individual sources should not give figures higher than 10%. If the figures are higher, then the level of similarity should be checked carefully to establish whether the copied material is posing a problem or not.

It is therefore proposed that each journal editor checks all submissions over a certain period of time to find out what is normal for that journal and/or discipline.

There are questions in RG which you may refer for further info.

https://www.researchgate.net/post/What_is_the_range_of_percentage_similarity_of_plagiarism_for_a_review_article

https://www.researchgate.net/post/How_much_plagiarism_is_allowed_for_the_acceptance_of_research_paper

As far as I understand no plagiarism is acceptable in any publication?

You should never used the exact words as appear in another publication, unless you put it in quotation marks and cite it.

Even if you are referring to your own work, rewrite it to say the same thing with different words.

I hear from one of Springer editor via webinar, there are < 20% acceptable in the introduction,and discussion section, also about 30% for theories.  

No amount of plagiarism is acceptable, this includes not citing other's work when it is paraphrased.

However, I disagree with Michael in one respect. Those of us who have to write 10 to 20 reports or so per year on a specific region must use boilerplate chapters for the history, prehistory, environment, geology, etc., changing them only to add a little something about the actual project area. There are only so many ways that something can be written. There are only so many ways to describe or explain something. If there are absolutely only two ways to write something, then is it plagiarism?

Reality check: plagiarism software picks up references and flags text in the section materials and methods, often indicating a problem where authors have effectively used the same methods in separate papers. You need to check carefully that you are not over-reacting. I am Editor-in-Chief of a journal and we use ithenticate but have a limit of about 30% above which we investigate in depth. The software is not yet sufficiently sophisticated to remove the need for human judgement.

Dear Fikrat, 

The acceptable percentage could vary depending on sources of the similar texts.

 According to springer guideline (see attached), editors should  Interprete reports and Read Percentages carefully. A high similarity percentage should be further investigated for:

Text from Many Individual Sources

How many individual sources are linked to this paper? The best way to check this is to use the Content  Tracking mode in iThenticate. A paper with single sources where each source shows similarity of only 1-5%, is generally showing no sign of potential plagiarism, provided that the majority of the paper is not made up of borrowed strings of material.

High Percentage from a Single Source

Some papers with a low overall similarity percentage can have a high percentage from a single source,  for example, the overall percentage is 15%, but 12% is coming from one single source. Generally, the  individual sources should not give figures higher than 10%. If the figures are higher, then the level of similarity should be checked carefully to establish whether the copied material is posing a problem or not.

It is therefore proposed that each journal editor checks all submissions over a certain period of time to find out what is normal for that journal and/or discipline.

There are questions in RG which you may refer for further info.

https://www.researchgate.net/post/What_is_the_range_of_percentage_similarity_of_plagiarism_for_a_review_article

https://www.researchgate.net/post/How_much_plagiarism_is_allowed_for_the_acceptance_of_research_paper

Dear Behrouz,

Many thanks.

Not plagiarism, but eye-opening about fair use in quoting media.

http://www.niemanlab.org/2016/06/1884-to-quote-300-words-from-the-new-york-times-in-a-book-two-authors-try-to-stand-up-for-fair-use/

Dear James,

 Its very interesting.

Many thanks

I am Editor-in-Chief of a major journal and we are also frustrated by the 'plagiarism detectors'. They are useful but need fine tuning (or interpretation and decision-making by humans!). We are alerted with scores >35% but, most of the time, the 35% is a sum of many items that are all

Dear Prof. Graeme B Martin,

Many thanks for your valuable contribution. The two points in last paragraph very important for editor with stream of potential, possible , or probable predatory scholarly open-access journals, so we may find in these journals the stealing work of other people .

One problem is that plagiarism detectors cannot distinguish between an archived draft and a final submission. So a plagiarism detector might indicate plagiarism ≧ 90% for an unplagiarized paper whose working copy had been archived by the author.

Another problem is that the presentation of research in the humanities often requires the use of block quotations, e.g. when a specific literary passage is reproduced so that it can be subjected to interpretation, analysis, and commentary, and so that the readers can judge for themselves whether the interpretation and commentary are credible without having to go to the source. Even though the block quotation may be properly referenced it is still identified as plagiarized material and can increase the plagiarism score by a significant percentage.