Ad Lagendijk Ad Lagendijk 9 February 2013

How to choose your key publications?

Tags: , , ,
Posted in High-impact journals, politics, Tips

The scientific community keeps on finding new ways to facilitate to judge scientists. The old-fashioned way of reading her papers, listening to her talks,  interviewing her for more than an hour, reading recommendation letters, and consulting colleagues personally takes way too much time.

List of Output becoming a book

Given the quality and ease of use of writing and plotting software the number of papers a scientist coauthors is exploding. The ease at which old email messages can be searched and the ease at which files onpersonal computers can be indexed and searched, leads to very long lists of outputs attached to c.v.’s. Conference proceedings, talks given at institutes, lists of successful research proposals together with the long list of publications makes the c.v. unreadable.


The old-fashioned way of judging scientists has two drawbacks: (i) it takes time and (ii) the judgement is done by the community itself. Especially the latter is a nightmare for managers and grant officers, who needs the power to control to upgrade their own c.v. Many non-scientists can count, so the counting of output has become an art. Publications are counted and citations are counted and lumped in indices likes the h-index. Classifying a scientist by just one number, her h-index, makes life very simple for the control freaks who want to control the scientists.

Key publications

More serious organizations have added a new way of classifying a scientist. The scientist is asked to report to a search committee or in a grant proposal the list of what she considers herself to be her best five publications. This requirements is a challenge for the scientist. What makes a publication a key publication. Here are some deliberations:

  1. Put in the list the publications you are most proud of.
  2. Put in the list the publications with the highest citation scores.
  3. Put in the list the publications with the most influential coauthors.
  4. Put in the list only your newest publications.
  5. Put in the list those publications that are most relevant for the proposal you are applying for.
  6. Put in the list only those publications where you are the last author.
  7. Put only papers there that have appeared in high-impact journals.

A number of these deliberations will lead to different lists. If you are a senior scientist and your most important papers are not of recent date your list will expose you as a scientist close to retirement. If you put only recent papers in your list, young committee members might not know that you are the author of a world-famous older paper.

My advice

The way I put together my five key publications is the following:

  • Two very recent papers in high impact journals.
  • Two papers I am proud of and have at least a reasonable citation score. No matter how old they are.
  • One paper in a high-impact journal with an age between five and ten years.
  • If your list is part of a collaborative proposal with colleagues also putting in their list of key publications there should be overlap, but not too much. So one shared publication with a co-applicant is advisory. Do not waste key publication space by different applicants listing the same publication.

I only list papers of which I am the last author.

Your way?

Our readers and me are curious how you put together your list of key publications.

- - - - - -
If you like this post why don't you email subscribe to our new posts. Or subscribe to our RSS feed.
  1. Mirjam

    9 Feb 2013 10:31, Mirjam

    I’d say it depends on the goal. If it is to get a tenure track position then you may want to pick publications that show that you have groundbreaking ideas, can work independently (so, not only 2nd author papers) and have certain skills (experiment, theory, numerical,…). If it is for a grant proposal you may want to pick papers that show that you are the best person to tackle that particular topic. If it is for a prize it could be only the highest citation papers, but that depends on what the prize is for. I think the list that you compile also says something about what you find important and in that sense it is a much better measure than a simple number such as the h index. Note that in spite of the obviously large impact God would not have done well with the current evaluation criteria: only one publication which is a book and not a paper, the wrong language, no references, no citations, no collaborators…

  2. Unregistered

    9 Feb 2013 10:32, David Stern

    Here in Australia we need to put 10 “career best” publications for ARC grant applications. I used the annual rate of citation to select a first draft of the 10. That gets some recent papers in as well as older ones. The I made sure both my Nature papers are in. The older one is only 19th by citation rate per year. And a couple more tweaks – taking out a couple of quite highly cited papers in weaker journals.

  3. bramvanginneken

    9 Feb 2013 15:43, bramvanginneken

    Do you really mean influenceable in point 3 or rather influential?

    What I like about the h-index is that you can simply look at those h publications, and that tells a lot about the researcher. So don’t use only the h number, but just use it as a handy tool to sort somebody’s publication list. I think most people will pick their 5 or 10 key publications from their list of h most cited publications, with some tweaks indeed, like David Stern does. Or, like David, using annual citation rates instead of the overall number of citations as the h index does, because that favors your older work.

    Taking out highly cited papers in weaker journals, why would you do that? In my opinion post-publication review (many researchers cited your work) is more valuable than pre-publication review (a very small number of reviewers allowed you to publish in their ‘top’ journal). I think the increasing emphasis on h-index is weakening the value of publications in ‘top journals’ (many of these publications are in fact never cited at all, even for the very high impact journals!) and that is a good thing. It will allow open access journals to thrive, eventually.

  4. Ad Lagendijk

    9 Feb 2013 22:24, Ad Lagendijk

    thank you very much. Of course it should have been “influential”.I have corrected this. Interesting analysis by you of the h-index. The point is not what is more valuable, but what will impress the reviewers of your proposal or the members of the search committee more.

    In my list of my key publications are certainly not my best-cited papers. But I agree that an h-index analysis (Web of Science or Google) of a scientist is a useful tool.

  5. Unregistered

    10 Feb 2013 0:19, David Stern

    Hi Bram – if I add the paper from Nature, I’ve got to take another one out. And it’s not like it has no citations. Actually, I just checked on Researcher ID and it is ranked #10 with 65 total citations, but it’s annual rate is lower as it was published in 1997. I think this paper will impress the reviewers more (who are an interdisciplinary panel) than a paper in B level economics journal that happens to have got more citations per year.

  6. Otto Muskens

    1 Mar 2013 23:38, Otto Muskens

    The UK government use a Research Excellence Framework” as a guideline for distributing funding. The idea is that every scientist selects his/her best four publications since 2008. A panel of senior scientists will evaluate all of these outputs, ranking each publication up to four stars. So how to select you four outputs? They emphasize that citations are not being used to select excellence, however the rules are deliberately kept very vague. The communis opinio is that a first author or last author publication is more important than a paper where the researcher is somewhere in the middle. Four papers on the same topic will count less than a broader coverage showing diversity of the researcher. It is emphasized that these panels of experts will not just look at the journal or number of citations. In practice this is a very expensive and enormous operation with panel members having to eyeball hundreds of outputs so one can wonder how thorough this assessment can be.

XHTML: You can use these tags:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>


By submitting a comment here you grant this site a perpetual license to reproduce your words and name/web site in attribution.

Subscribe without commenting