Checking for originality: Crossref Similarity Check
Download as PPTX, PDF1 like941 views
Similarity Check is a similarity screening service that checks submitted manuscripts against over 59 million publications from Crossref members, 105 million publications from other partners, and 60 billion web pages to generate a report for editors. Between October 2016 to October 2017, over 4.7 million manuscripts were screened using the iThenticate software. For publishers to use the service, they must register content and assign DOIs so it can be indexed in the Turnitin database. As-crawled URLs in DOI metadata are used by Turnitin's crawler to index content. Similarity reports help publishers look for issues like plagiarism, self-plagiarism, unattributed use of others' works, or misrepresentation. Publishers invest significant
Checking for originality: Crossref Similarity Check
- 2. What is Similarity Check? • Not plagiarism detection • It is a similarity screening service that uses the iThenticate text comparison software that was developed by Turnitin • What are papers checked against? • 59.1 million items from 1,271 Crossref organizations participating in the service (over 1,500 distinct members) • 105 million items from other content partners like Pearson, McGraw Hill, Cengage, EBSCOHost • Over 60 billion web pages archived back nearly a decade • Then provides a report for editors to evaluate
- 3. Good and growing adoption Total number of manuscripts scanned in iThenticate between 1st Oct 2016 - 1 Oct 2017 was 4,720,098
- 4. Setting up the service • Must be registering content and assigning DOIs in order to start using the service • Content must be indexed to be added to the Turnitin database • Indexing is done via as-crawled URLs. These must be present in your Crossref DOI deposit metadata • Function to deposit as-crawled URLs available in OJS 2.4.7 and above
- 5. What are as-crawled URLs? • Link used by Turnitin’s crawler service to identify and index the content associated with your DOIs • The as-crawled URL points the location of the full- text content, either PDF or HTML, associated with the DOI. • New DOIs the URL an be included as part the deposit metadata • Existing DOIs maybe updated
- 7. What issues are publishers looking for • Poor, missing, or incomplete references (can be fixed) • Self-plagiarism/text recycling (can be fixed) • Unattributed use of parts of another person’s work • Submitting another person’s work as your own • Is the author attempting to mislead/misrepresent?
- 8. How are people using it? Publishers putting time and effort into their policies and workflows: • Resources (staff & time) • Cost • Workflow • What will you look for & how will you look for it? • Education • Follow-up actions
- 10. Cost • Administrative fee = 20% of Crossref member fee • Document checking fees paid annually to Turnitin • Each publisher has their own login • More cost effective than joining Turnitin directly
- 11. Demo
- 12. Questions?
Editor's Notes
- #2: CR offers a number of services for publishers to use to enhance their publication record I’m going to discuss one of first services we added several years ago, Similarity Check.
- #3: SC is not a plagiarism detection service. It is a similarity screening service that uses the iThenticate text comparison software that was developed by Turnitin. SC publishers are able to check and compare their documents against database of full text articles.
- #4: Checking both English and non-English content. Only checks the language in which the content is written, does not translate content items, more being added all the time.
- #5: First Step! Crossref will tell Turnitin precisely what to Crawl pointing only to your DOI-identified content. Once indexed Only crawl when new content added or updated As-crawled URLs must be present in your Crossref DOI deposit metadata before you may begin to use the iThenticate service.
- #6: An as-crawled URL is a specific crawler-friendly link used by crawler services to index content. turnitin needs the as-crawled URL for the content associated with each DOI in order to index your content as part of your Similarity Check membership agreement. Even if your as-crawled URLs are the same as your DOI resource URLs, you still need to enter them separately as specific as-crawled URLs. This allows the service to scan the text against submitted manuscripts. Does not provide outside reader access to content.
- #7: Shows the % of text that is similar. Check where the match is from. The similarity percentage can be misleading if not interpreted properly. Some text may be similar, such as properly cited references; standard scientific description – such as the Materials and Methods used in an experiment or research You can exclude certain or refine the % threshold.
- #8: In order of severity, Self-plagiarism is the use of one's own previous work in another context without citing that it was used previously.
- #9: Using Similarity Check as a TOOL. Different places in submission process Three common places where a publisher might want to do plagiarism screening 1) On submission 2) At some defined point in the review and editorial process - - varies widely form publisher to publisher 3) Just prior to acceptance And we have Sim Check members taking each of these approaches - trend moving towards screening on submission.
- #10: Have a policy in place before first before implementing the service.. How are you going to deal with evidence of plagiarism. Publishers creating their own resources to deal with the service, and using COPE. Committee on Pub Ethics