CrossRef Reports to Publishers: “Issues” and “Conflicts” are simpler to resolve than you might think

Publishers (and their deposit agents such as HighWire) get regular reports from CrossRef that spark some questions.   The reports are emailed, and the email subjects have ominous-seeming titles, warning of danger, or at least error.   The subjects of the emails are these:

  • “DOI issues”, with an attachment of “DOIerrs”
  • “Conflict Report”

Publishers shouldn’t panic.  But you should open the reports to see if the situations they describe are acceptable to you and your readers, and to your authors.   Here’s what’s going on:


“DOI issues” is a report that shows that some “DOIs” with your publisher prefix have been entered on a web site or in a web browser, but did not resolve to any registered DOI.   In other words, something was incorrect in a URL.  If you open the email attachment and look at the “DOIs” that are listed, you might often say,

  • “Hey, that’s no DOI of mine! My DOIs don’t look like that.” or
  • “That’s my DOI, but it has a period at the end” or
  • “That looks like my DOI, but it has one too few characters”.


In most cases, somebody typed a DOI — or perhaps cut and pasted it — and didn’t do a good job.   Digits were transposed or missing, or the letter O was put in place of the number zero perhaps.

In a few cases you might see a legitimate DOI, and wonder what went wrong.  There are several possibilities, but one thing to check is the “reported date” and see if the article had been published then.  Sometimes publishers provide a DOI to the media, and the reporters try and use it while an article is not yet formally published.  This will result in a DOI “issue”.

If you see no referring URL in the report, it may mean that the user typed it in by hand; but it could also mean that they were browsing privately.   Hand-typed URLs are hard to type accurately — like dialing an international phone number.

The Fix:

If the referrer in the report is a significant source of traffic — and certainly if it is a page on your own site! — you should go to that URL and find the reported DOI on the page (using the browser’s Find function).  You can then request the owner of the page please correct the DOI.  These referring pages can sometimes be blog pages or news pages.  Be sure to look not only at the text on the page, but at the hyperlinks behind some of the text — the DOI might be inside a hyperlink not visible as page text

Sometimes there will be a simple and obvious fix — e.g., someone included the period at the end of the sentence in their DOI hyperlink.   That is a pretty easy mistake to make, and to correct, if you can get to the owner of the webpage with the mistake.  Fortunately, the CrossRef report can tell you how many times a particular page has been used — e.g., is it a news item in Science or Nature? —  so you can see how much effort it is worth to get it fixed.   Usually, the more high-traffic sites are also the ones that will fix an error quickly.

Want more details?

For more information about this report, see the CrossRef documentation:


The Conflict Report is intended to let you know that it appears that you’ve got a duplicate article — based on CrossRef’s analysis of the deposited metadata.   Of course, CrossRef can tell only so much from the metadata.  A duplicate article is an article for crossref is an article with the same title, volum, issue, date and page.


You might wonder how multiple articles could end up with the same volume, issue, date and page.

By far, the most frequent cause is articles whose titles are short and common: “Correction”, “Erratum”, “Editorial” and sometimes things like “Book Review” or “Editors’ Choice”.  Of these, typically only Correction and Erratum will have multiple “articles” on the same page, so that is the most frequent situation.    If you do not fully title your articles — e.g., give a correction the title “Correction” rather than “Correction to Smith et al, 2016” — then this is the outcome

Another outcome of using generic and short title — beyond the CrossRef report  complaining about it — could be that search results in Google and Scholar will be unhelpful.  Try a “fielded” search with the common term, e.g.,

site:<yourjournalsiteurl> Correction

e.g., Correction

This alone is enough of a reason to go beyond a generic title for things like Corrections, Editorials, and the like.

It is also possible that one article has accidentally been assigned two different DOIs.  This is very rare.

The Fix:

There really is no fix other than to change your practices to provide complete information in the article title.  This may require corresponding changes at HighWire, if you have had us build your site to work with generic titles.

Want more details?

If you publish biomedical information might want to look at PubMed’s recent “best practice” advice on titling such items, copied below:

We want to take this opportunity to make the following requests related to submission of citation data for errata and retractions:

1. Please add “Erratum” (or “Correction” or “Corrigendum”) to the beginning of the article title if the title of the erratum notice is exactly the same as the title of the article being corrected. This change will help PubMed users differentiate between the publications.

2. When submitting citations for errata and retractions, please use the PublicationType values “Published Erratum” or “Retraction of Publication” as appropriate.

3. We encourage data providers to include the PMID, DOI, or PII of the original PubMed citation when submitting a citation for an erratum or retraction notice. When this information is included in the XML citation data, the link between citations will be displayed when the citation for the erratum or retraction notice becomes available in PubMed. Please see our Help for tagging instructions:

3 thoughts on “CrossRef Reports to Publishers: “Issues” and “Conflicts” are simpler to resolve than you might think

  1. On the other hand, Crossref is reporting huge problems with the funder metadata they are getting from publishers. With funder metadata for almost a million articles submitted, Crossref reports that over 50% of these are useless. The biggest problem by far is that the submission does not include the funder ID. This may be in part because there are over 10,000 funders in the registry, so it is laborious for the publishers or their vendors to find the ID for every funder identified for a given article. In any case at this point the funder ID process simply is not working. See

    I have been tracking this problem for a long time, in my weekly newsletter Inside Public Access. Moreover there are a lot more publishers gearing up funder identification for the first time, so this is not about to go away.

    Does Highwire do funder identification for its publishers? What is their experience?


    • David,

      I just heard that what you report was a discussion on the CHORUS Working Group.

      At HighWire, for our BenchPress manuscript-management-system customers we have a widget that supports validation. But for others who supply us directly with tagging, we don’t validate that a funder ID is in the funder table. I think this type of validation needs to be done upstream, where the people who know who the funder is can make the right choices. I wonder if CrossRef (or CHORUS) will be producing an “error report” that points out where the funder IDs do not map to anyone in the ID table?

      It is a good thing for the development of ecommerce online that in order to buy something you don’t have to put in both a bank number AND a credit card number.
      Imagine a situation in which there were 10,000 credit card companies, and everybody had a different numbering system for their cards. I guess we’re lucky there are only a few credit card companies, and (apparently) based on the credit card number you can tell which company “owns” that card.

      Have you ever tried to pay by check online? You have to put in a lot of numbers. I’ve rarely done it, but it usually involves me looking up pictures that tell me which numbers are called what. Maybe that is what authors are going through the first time or two they submit through a system that requires funder ID and grant ID.

      I wonder how much this is a startup problem: authors getting used to systems, getting to know the accurate names of their funders, etc. Remember all the confusion about PMC deposits years ago? Has that settled down, from authors’ points of view?



      • My guess is that CHORUS is ahead of the curve on this, since ACS and APS pioneered funder identification as part of the CHORUS startup. But supplying the Funder ID# is not an author problem, as it is something the publisher or submission system vendor does. The author just has to get the funder names right, which apparently many do. Someone should be working on name recognition software that assigns Funder ID#s.

        I am sure it is a startup problem but it will take several years for everyone to get up to speed. CHORUS members have a year to implement funder identification and many have just joined or have yet to join.

        In my view CrossRef (and Elsevier, the funder list owner) overreached when they decided to try to include every funder in the world, in the startup phase. I warned them about this early on. Just getting the US Federal funders right is a challenge, because of the highly distributed nature of Federal funding. Hundreds of different offices and programs fund research. It’s a zoo, one I have studied for many years.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s