13

I have received an immediate reject notice from the journal ACM CSUR mentioning that my written survey had 13% similar materials with respect to the online available sources. Passing through the instances one by one, it sounds to me that the majority are just the text of the references to the citations, e.g., "Authors et al. [first_author et al. Year]", some standard terms such as "performance", "tuning", and many standard ACM sections such as the ACM Copyright section, etc.

P.S:

1- To be noted that I am aware of plagiarism policy and I strongly support it. However, the fact that we are dealing with a survey journal instead of a regular computer science experimental research paper made me confused on the outcome of the similarity check and the immediate rejection.

2- The application that does this check is called "iThenticate".

Had anyone experienced the same issue ? Normally what do we do to proceed in this status? Does ACM have a fixed threshold on this similarity score?

| improve this question | |
  • 6
    13% of a paper is about 3 words for every single line of text. Are you sure it's really just "Authors et al. [...]" and standard terms? – Wolfgang Bangerth Oct 31 '16 at 19:31
  • @WolfgangBangerth, My journal has 25608 words and there are 60 sources reported to which I had similar phrases with. My top 6 which are above 1% are the one I mentioned. The other small sources having 20-40 words are taken from papers, mostly having similar clauses of 4-5 compound words. But really, these are standard sentences we use in my field and my written language hasn't changed much from the papers I published at IEEE or ACM before. – Amir Oct 31 '16 at 19:43
  • I have recently experienced this issue. What did you do in this case? – user5155835 Oct 3 '17 at 17:45
  • @user5155835, I revised the whole text by considering the provided similarity document and to avoid any future issues. – Amir Oct 4 '17 at 18:05
  • @Amir how many months has it been under review? – user5155835 Oct 5 '17 at 4:51
5

I cannot find anything about the use of the ithenticate in the ACM CSUR instructions for authors. The ACM magazine Inroads says:

Please note that ACM Inroads as well as all ACM publications expect new, unpublished material, unless specifically reprinting an item from another publication with attribution. In support of this goal, ACM Inroads uses iThenticate to develop a similarity index for all submissions. As a first step after an initial submission, iThenticate compares the text to published material (at least 43 billion web pages and over 130 million content items). It provides editors with a set of similarity indices and helps them identify original material, including prior publications of authors.

This suggests that there is not a hard cut off, but rather a set of metrics and comparisons that the editors use in making a decision. While it is possible that the ACM system is poorly tuned, this seems unlikely. In my experience, a 13% similarity is quite high and is often indicative of issues.

While it is not cheap, you can purchase a single paper use of iThenticate. It appears that iThenticate is owned by TurnItIn which many universities have access to. It is not clear what the different systems do, but it may help you identify what is being flagged up.

| improve this answer | |
  • 2
    @Amir I would expect some human intervention. Despite tuning the systems, sometimes things sneak through (e.g., block quotes or standard questionnaires). – StrongBad Oct 31 '16 at 20:11
  • 7
    Turnitin is designed for use on student assignments while iThenticate is designed for use on research papers intended for publication. Turnitin's database has lots of student papers, while iThenticate checks against published works. iThenticate frequently gives false alarms for standard phrases like "Without loss of generality, we can assume that." In my experience, 13% is towards the upper end of what I would expect from a paper that didn't include plagiarized texts, and as an editor I'd be examining it carefully. An editor really needs to evaluate this carefully. – Brian Borchers Nov 1 '16 at 2:12
  • 2
    @Amir I have started to rephrase and rewrite certain blocks of text to reduce the percentage This really sounds an awful lot like you do not fully understand what plagiarism is. – StrongBad Nov 2 '16 at 16:29
  • 1
    @StrongBad, We are talking about a SURVEY paper here as I mentioned in the question and this is exactly what I have been guided by the email received from the editor in chief. – Amir Nov 3 '16 at 16:51
  • 2
    iThenticate and Turnitin are different services offered by the company iParadigms and they appear to use the same databases. Even a survey paper should quote properly, @Amir. Indeed, one of the weaknesses of iThenticate is that it does not recognize citations properly and will flag references as being text overlap. Did they send you the report? Just rewriting so that iThenticate won't scream will result in an entirely unreadable paper. I would start afresh and be aware when I am summarizing (and making sure I reference) and when I am quoting (and of course reference here, too). – Debora Weber-Wulff Dec 4 '16 at 11:44
3

Contact an editor and explain it like you have explained it to us.

Unless this Ithenticate thing is new, it is likely that they have already encountered this problem before and they know what to do.

| improve this answer | |
  • Thanks for the answer. I have revised the text and took into account each of the match instances. This time I believe I could pass the initial phase as I haven't received any immediate rejection so far. – Amir Nov 18 '16 at 10:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.