Dear Ahmed Taha , text overlap is not the same as plagiarism. Plagiarism is the re-use of the wording or ideas of someone else without attribution, which is never acceptable. Plagiarism is the presentation of  work of others as your own. You need to aim for 0% plagiarism and here should be 0% plagiarism in any professional article. 

You may quote others briefly in quotation marks, and methods descriptions can usually be repeated if the source is clearly cited. Authors can sometimes re-use parts of their own wording providing they have permission from the copyright holder (easy if they retain copyright) and they are transparent about the re-use. Beyond that, you need to write articles entirely in your own words.

Text overlap percentage depends on how the comparison is done, and should only be a guide for reviewers/editors.  The plagiarism level usually acceptable is 25 - 30%. But this limit is because of the natural requirement to quote something, cite a reference, or include a part of results from previous researches and build upon that.

This limit is not for directly ripping off lines from a paper or research work or any other source. When I do comparisons, I exclude direct quotes and references as I expect there to be repeated text -- that is supposed to happen!

If you use the words, diagrams, graphs, etc.  of another, then those items should be quoted or set off in some way, with either a citation reference or a footnote showing their origin.   References, by their very nature, show their association with the work of others and cannot be plagiarism.   (However, the references must be real, and you had to have actually read them!)

These days good journals accept a manuscript containing only 15 percent of plagiarism.

Dear Ahmed Taha , text overlap is not the same as plagiarism. Plagiarism is the re-use of the wording or ideas of someone else without attribution, which is never acceptable. Plagiarism is the presentation of  work of others as your own. You need to aim for 0% plagiarism and here should be 0% plagiarism in any professional article. 

You may quote others briefly in quotation marks, and methods descriptions can usually be repeated if the source is clearly cited. Authors can sometimes re-use parts of their own wording providing they have permission from the copyright holder (easy if they retain copyright) and they are transparent about the re-use. Beyond that, you need to write articles entirely in your own words.

Text overlap percentage depends on how the comparison is done, and should only be a guide for reviewers/editors.  The plagiarism level usually acceptable is 25 - 30%. But this limit is because of the natural requirement to quote something, cite a reference, or include a part of results from previous researches and build upon that.

This limit is not for directly ripping off lines from a paper or research work or any other source. When I do comparisons, I exclude direct quotes and references as I expect there to be repeated text -- that is supposed to happen!

If you use the words, diagrams, graphs, etc.  of another, then those items should be quoted or set off in some way, with either a citation reference or a footnote showing their origin.   References, by their very nature, show their association with the work of others and cannot be plagiarism.   (However, the references must be real, and you had to have actually read them!)

Following

As Beemnet says: You must aim for 0% plagiarism and there must be 0% plagiarism in any research article. 

Agreed with Mr. Kassahun statements. Please read that and proceed accordingly.

Interesting question and valuable responses

Plagiarism is not permitted at all. 20-30% text overlap is permitted

The similarity index for good journals is pegged at 15-20%. Exceeding this range will render the paper not publishable. Best regards

Good contributions

15%-20%

It varies

Interesting question....following.

It all depends on the journal ethics and plagiarism policies. Before you submit your article, read the scope and publication ethics of the journal. High impact journals are very serious about quality publications. Most of them uses plagiarism tracking software. For example your article must meet the bench-making between 15-18% then it can be concluded as originality. I think your article will be arrived to 0% because you must also cite sources. These high impact journal are accredited by COPE- Committee of Professional Ethics because they all must be aligned to their requirements of such body. finally, I preferred 15-22% variance in all my articles but some times it is not possible to arrive to such a bench-making.

Maximum 20%.

Kindly visit these links:

https://www.researchgate.net/publication/311665427_Journal_Impact_Factors_Researcher_h-index_How_to_improve_research_quality

https://www.researchgate.net/post/Trusted_and_indexed_Journal_Impact_Factor_2017_Updated_by_Thomson_Reuters_Journals_Citation_Report

https://www.researchgate.net/publication/313423128_CiteScore_vs_Impact_Factor_narrated_by_Asst_Prof_Dr_Emad_Kamil_Hussein

https://www.researchgate.net/publication/320255345_The_Iraq_Virtual_Science_Library_IVSL_by_Asst_Prof_Dr_Emad_Kamil_Hussein_October_2017

https://www.researchgate.net/publication/313750336_IF_YOU_WANT_TO_by_Asst_Prof_Dr_Emad_Kamil

Regards,

Emad

I think less than 10% is suitable for all journals.

10-15% for the entire document to allow for direct quotes, references and general formula. In the text, the allowable percentage would be 0%

Zero.

20%

Should be less than 20%. But its depend upon the plagiarism policy of publisher.

As Beemnet says: You must aim for 0% plagiarism and there must be 0% plagiarism in any research article. 

As I indicated earlier, read the policies of the particular journal before you starting uploading your article or book chapter. As an editor-in-chief of my journal I put all articles through iThenticate plagiarism tracking system (COPE) and found more than 20 % similarity, I forward both the article and report not to be judgmental to develop the author's work for corrections. I preferred below 15% as a bench-mark for all journal article.

It all depends on the journal ethics and plagiarism policies. If there is no information about it on the journal, I think for most of editors; 10-12% for the entire document to allow for direct quotes.

Thanks so much to all for sharing ur experience.

Good question and interesting discussion! It seems that there is no consensus. The issues about plagiarism must also be taken into account in case of using personal work without citing the source (a previous article or a theses). Journals can have different policies related with plagiarism, however in a dissertation or a theses this ethical problem requires vigilance. In my university when students deliver their dissertation or theses they have to sign a declaration indicating that there is no plagiarism in their work. So the percentage of plagiarism allowed is zero. All the students work is examined with an antiplagiarism software. Even after the defense if a plagiarism problem is detected the degree (MsD or PhD) can be annulated.

I think less than 20% is suitable

Yes, I agreed with all the RGs inputs of 20% as a guide. But, if the similarity is more than 20%, what now, and how are we dealing with those parts of the 20%. Any suggestions to correct to it?

Most institutions I have work with provide guidelines and encourage students not to exceed 20%, however if you exceed 20% due to references and quotations or your study is based on law topic you would not be penalised. Moreover, as pointed by Maria Joao Loureiro above citing your work makes a lot of difference calculating percentage of similarity in your research.

It is highly appreciated for all your inputs on similarity of work published. For our students, whether undergraduate or postgraduate, clear guidelines are provided (20%), using Turnitin. For academics, all research outputs, the iThenticate software programme is compulsory. I tried to keep all my articles below 10% (this is sometimes difficult when your report empirical data). It is more easy when I write critical review of argumentation papers or conceptual papers. Each paper I published I know it must below 20% similarity (included references).