Did you receive spam because you published a paper?

Brendan Thomas has published an interesting paper that looks at author email addresses in the PubMed database of biomedical literature. Email addresses of first authors have been added to PubMed since 1996, and they can be retrieved via the standard web interface or automated software. This makes PubMed an excellent place to find the email address of an academic author, but also shows that PubMed is very vulnerable to email address harvesting.

The problem is not limited to PubMed; this is also an issue with many scholarly journals. Email addresses are used as contact information in scholarly papers, and they are commonly displayed on journal webpages. You can harvest email addresses from NEJM, Science or PLoS webpages, and you don’t even need a subscription. JAMA is hiding the email address, and Nature is providing an author contact form. Both journals provide the email address in the PDF.

Most journals provide the postal address of corresponding authors. Journals added email addresses in the 1990s, and this of course has become the preferred method of communication among scientists – when was the last time you received a postcard for a reprint request? Unfortunately we have long learned that it is no longer safe to make email addresses publicly available – more than 90% of email traffic is spam.

Journal publishers have to rethink their policies regarding contact information of their authors. And authors should demand from journals that their email addresses are treated with more respect for privacy. There are better ways to provide contact information for corresponding authors, and I don’t mean a link to their Twitter account or LinkedIn profile. Journal publishers should create author profile pages that not only list all publications of a particular author, but also the relevant contact information. There should be contact forms instead of plain email addresses, and authors should be able to control what information is displayed in their author profile. Author profiles of course can be further extended in many ways, from links to publications with other publishers to author-level metrics.

Disclaimer: I sit on the Board of Directors of the Open Researcher & Contributor ID (ORCID) initiative which aims to help solve this and related problems.

Thomas, B. (2011). E-mail Address Harvesting on PubMed–A Call for Responsible Handling of E-mail Addresses. Mayo Clinic Proceedings Mayo Clinic, 86(4), 362. doi:10.4065/mcp.2010.0817

Related Posts Plugin for WordPress, Blogger...
This entry was posted in Conferences, Interviews, Presentations, Recipes, ResearchBlogging, Reviews, Snippets, Thoughts and tagged , . Bookmark the permalink.

8 Responses to Did you receive spam because you published a paper?

  1. A contact form on a website is no replacement for an email address — especially when the maintenance of those pages is dubious at best.

    When readers take their own copy they should get inline information, not just a link to a profile; do you envision that being dynamically updated?

    For examples of existing author profiles, ACM is doing as good a job on this as I’ve seen.

  2. Jodi, I don’t go into this in the post, but I think that a universally used identifier is required for this, and of course I hope that it will be ORCID. This identifier can then also be used in the PDF of the paper and in the metadata that are sent to PubMed and other places. The journal author profiles could then be built using the information in the ORCID profile. The ACM Digital Library will interact with the ORCID system (Introduction to ORCID, slide 13).

  3. People have been using antispam corruptions of their addresses for some time, such as myself@thisdomain.com becoming myself [at] thisdomain [dot]com, or myself[DeleteThisBit]@thisdomain.com
    Simply journals need to do the same, including in respect of the published pdfs. Authors should send in pre-corrupted addresses, but then some journal systems reject anything that isn’t a valid email. That’s the one change that is needed here.

  4. Martin Fenner says:

    Robin, this is of course an alternative strategy. I prefer author profiles, as they can also contain other useful contact information, e.g. the departmental webpage of the research group, or even the Twitter handle of the corresponding author.

  5. Carlos Rocco says:

    Hi. I Have a question, not sure that this is the right forum. I want to harvest emails from pubmed to advertise a short course. Despite the course has a fee, our intention is not to make a profit but rather to finance costs and scholarships. Moreover, we are planning to include a narrow profile of investigators based on their areas of interests, and we are probably writting a different (personalized) email for each one (we are not bulk-mailing). We certainly have no experience organizing an international course and maybe you can answer if this is a valid/effective way of diffussion. We do not want to bother any authors, but we think the ones that we are mailing would be interested. Perhaps you can suggest a better alternative. Thanks.

  6. Carlos, I would never advertise something via email.

  7. Carlos Rocco says:

    Tough, i saw it comming thoug. I think, from your post, you believe that : a) it would be a privacy violation, and b) author will not read it. I think, also from yor comments, that a better option would be to mail the departmental webpage of the research group. I will also try some specialized blogs. Any other suggestions? Thanks

  8. Neil McKenna says:

    I think e-mail addresses should be removed from PubMed and PubMedCentral. I get vast volumes of spam from companies I have never heard of, and I strongly suspect I ended up on these e-mail lists through my e-mail address being lifted from PubMed and PMC. If a scientist really wants to contact an author they can access the full text of the article through their institutional subscription, or by going to the library and looking at the paper version. This obnoxious, intrusive practice and the people who indulge in it must be stopped.