CrossMark: There is no final version

Editor's Notes

• #2: I ’ m going to be talking to you today about CrossMark - a new service from CrossRef that fits perfectly with the theme of the seminar today. CrossMark will provide scholarly publishers a better way of dealing with corrections and retractions and it will give readers of online journals better information about what they are reading.
• #3: By way of background to why we set about developing CrossMark I ’ d like you to consider a few issues. The first is that “ content changes ” and to often scholarly content - in particular journal articles - are seen as being preserved in amber - fixed and unchanging. But it ’ s not as simple as that....
• #4: Many things can happen to content after it has been published. It can be corrected, enhanced, retracted or even withdrawn. This has always been the case, and it has been and still is - in spite of the internet - the publisher ’ s responsibility to correct and update the literature that they publish.
• #5: It has always been the case that content can change after publication but things were different in the print era. Here ’ s a photo that a colleague of mine took when he was visiting a publisher not that long ago - it ’ s a correction notice for a book posted on a bulletin board - which is an archaic and pretty ineffective way of trying to let readers know about an error in a publication - but makes the point that mistakes happen and publications need to be updated and amended.
• #6: There is also the big issue of trust for scholarly literature. Changes in content, if handled correctly, can enhance trust but if they aren ’ t handled well, they can undermine trust. There has been quite a bit in the news lately about how scientific fraud may be on the rise, with suspected increases in plagiarism and other types of misconduct. An some studies have confirmed this. A news item in Nature last year confirmed the numbers and highlighted a 10-fold increase in retraction notices which far outpaced the growth in the number of articles.
• #7: But as I said we ’ re talking about more than just retractions. Corrections and corrigenda are more common, and in the online world there are growing opportunities to enhance content, perhaps by adding source data or supplementary material after publication of the original article. And there is no way to easily assess whether corrections, updates and enhancements are going up.
• #8: So two main points that CrossMark is addressing - First: it ’ s time for the idea of the “ final version ” of an article to rest in peace. There is no final version for content. Readers, and often publishers themselves, have a mindset that once the “ version of record ” is published that ’ s the end of the story and the end of the publisher ’ s role. In all the recent debates about whether and how scholarly publishers add value I didn ’ t see any mention of the role publishers have in providing ongoing maintenance and stewardship of content.
• #9: Secondly - if content changes as we know it does, then readers need to know. It could simply be that an update adds new information to an article, but it could potentially be more serious, with corrections to information that could alter follow-on research or even treatment. So it ’ s important that this information gets out there, and notices on bulletin boards aren ’ t a good way of doing that.
• #10: Of course publishers inform readers when there ’ s an update and nowadays they do so electronically, but there ’ s no standard way of doing this, and it ’ s not always easy to spot updates. This article in Science has a correction, and as it ’ s flagged in red text you ’ ve probably spotted it over here on the left.
• #11: But this one is different. There ’ s nothing in the left of right hand columns, but instead this publisher has chosen to site the correction up at the top of the article here and it is very prominent.
• #12: And what about this one? Nothing obvious at the top of the page, or in the tool bars on the right...
• #13: ...but if you scroll down the page - below the fold - here ’ s a correction located under the “ related articles ” heading. This is harder to find.
• #14: Which leads to a second problem, which is that there is often more than one version of an article available. Here we have an article from the Journal of Surgical Research which was retracted because it was found to contain plagiarized material. On the publisher ’ s site it ’ s flagged pretty clearly as retracted up here in the article title...
• #15: If you search for this article in Google Scholar, however, the publisher ’ s site isn ’ t the first to appear - in fact it ’ s the fourth listing. And there is no indication here that the article has been retracted.
• #16: The first result is an information sharing site for doctors where someone has posted the abstract, and there ’ s no mention of the retraction....
• #17: The second is PubMed, and the retraction has made it on to the Pub Med copy, although it ’ s not as obvious as it is on the publisher ’ s site - it ’ s not part of the article title but a separate link below.
• #18: But what if you ’ d come across the abstract somewhere else? Maybe through CiteULike, where again there ’ s no mention of the retraction.
• #19: Or there could well be a copy in the author ’ s institutional repository... With all of these options there ’ s a reasonable chance that the reader isn ’ t necessarily going to see the correction or retraction that the publisher has issued.
• #20: A huge problem is PDFs - what if I download a PDF to my computer when the article is published and then open it up a few months later after a correction has been made - how do I know if there ’ s been a correction?
• #21: These are all problems that CrossMark addresses.
• #22: So CrossMark. At its simplest it ’ s a logo that publishers will apply to content that they publish. When a reader clicks on the logo they will quickly and easily be able to tell: The best way to explain it is to show some examples.
• #23: I ’ ll start with the most useful common scenario. We ’ re looking at a PDF from the Journal of Applied Crystallography. This came from my hard drive. Or I downloaded it from the author ’ s web site. Or was it my university ’ s institutional repository? Maybe somebody emailed me a copy? No, wait, I think this was from my Mendeley account. And when was that? At any rate, you see there is a CrossMark logo in the upper left corner. Providing I am online, when I click on the logo it will pop up a webpage...
• #24: with a pop-up dialogue box giving the latest status. This is what most people will see - confirmation that the document is up to date, the CrossRef DOI link that will always point to the publisher-maintained copy, and a link to the publisher ’ s policies. There are no updates. This time.... (click) And the box also tells the reader that Future updates - if any - will be listed below, so getting them used to the idea that if changes happen, this is where they can find them. But what if there had been a correction?
• #25: Here ’ s another PDF. (For the record, this example is a mock-up from the fictitious Journal of Psychoceramics.) The CrossMark logo in a different place (click), though it works the same way. I just click on the logo. ..
• #26: Here, clicking on the logo brings up the same CrossMark dialog box... but I discover that there was a clarification issued for this document. It gives a link to the correction. All I need to do is click on the CrossRef DOI link (click) to go to the update.
• #28: Of course, these work the same way on publisher ’ s HTML pages. (Click) You see a real article from The Proceedings of The Royal Society B on Royal Society ’ s website. The CrossMark logo appears just above the article title here.
• #29: On a web page, rolling my mouse over the logo brings up a text box that says “ Click to get updates and verify authenticity. ” And then when I click on the logo...
• #30: I see the CrossMark dialog box. All is well. (click) One of the pieces of information in the box is a link to the publisher ’ s CrossMark policy.
• #31: As a requirement of participation, publishers create a CrossMark policy statement that explains how they handle corrections, retractions, or other updates. CrossRef doesn ’ t have anything to say about what those policies are--just that a publisher must describe them in a policy document.
• #33: This example (also real) is from the International Union of Crystallography. I click...
• #34: ...and see that it has a correction. Like before, I can click on the link to the correction. The publisher has used CrossMark to do something else in this example. (click) It has provided optional publication record information about the article.
• #35: This is where you can find additional metadata about the piece of content. The publisher decides what to put here and can use these fields to define publication practices. They may choose not to populate this information. If they don ’ t supply an additional metadata the tab simply won ’ t show. The fields are defined and labelled by the publisher, and there can be as many or as few as necessary.
• #36: These are a few of the other possible pieces of information that have come up when talking with publishers. CrossRef isn ’ t going to advise on what publishers should display in the record box, but we expect that communities of interest may develop guidelines or best practices within different areas. There ’ s already a group of publishers discussing how best to display funding and grant information, for example.
• #37: CrossMark data will be freely available and machine readable and queryable, so could potentially be used in search results to flag content that has status verification and possible additional information, although this is something that we haven ’ t discussed at any length with the search engines just yet. We have had some conversations with librarians about using CrossMark data to populate link resolvers by pulling back relevant information, and also with other third parties such as bibliographic management systems who might be able to pull status updates into users reference lists and personal libraries. Another possibility is for A&I services to mark records with the CrossMark logo.
• #38: The first thing is that you have to be a CrossRef member in good standing. But CrossMark is an optional service, so there is no requirement for CrossRef members to participate. CrossMarks can be assigned to any piece of content that has a CrossRef DOI. This can include ahead-of-print or early release copies that the publisher has made available and is committing to maintaining. CrossMark logos should not be applied to author ’ s copies or any other pre-print that is outside of the publisher ’ s control and will not be maintained. So by joining CrossMark you are agreeing to maintain your published content, and to keep it ’ s associated CrossMark metadata up to date. We also have some guidelines on displaying the CrossMark logo.
• #39: There is no requirement to add CrossMarks to backfile if the publisher doesn ’ t want to do so.
• #40: You need to display the CrossMark logo on the article ’ s CrossRef DOI response page, ideally situated as close as possible to the article title, and outside of access control. On the CrossMark pilot site you will find information on the code that can be inserted into HTML pages to add the logo and link it to the CrossMark dialogue box. The CrossMark logo should also be added to PDFs. Black and white versions of the logo are available if required.