A PI started a study and collected some data. The study was originally meant to evaluate the effectiveness of a treatment. After a year or so of collecting data, the PI decided to do some analyses. When I looked at the design of the study, I informed the PI that this design will not be able to provide any meaningful evaluation of effectiveness.

Some time later, I came across a novel method to analyse this data and tried it out. The results were quite interesting and with the PI's agreement, I decide to try and publish it.

My question is this: Since the PI was not involved at all in the conceptualisation of this novel method and did not contribute in the interpretation of the results, should he be granted authorship? According to ICMJE guidelines, the answer should be no but I am unsure.

Addendum: I would like to add, while thinking about this problem, I thought of big studies like the Framingham Heart Study, the Global Burden of Disease study. The PI of these studies are not automatically authors on every paper that came out of the data they collected. Would this be any different from my situation?

  • 14
    It is still PI's data. Without his data you would have no research. – Alexandros Jun 10 '14 at 3:06
  • 2
    Two questions: (i) What is your relationship to the PI? Are you his student or postdoc? Is your work on this paper clearly part of the scope of the funded project, or is it more ancillary? (ii) What was your understanding with the PI about the nature of the collaboration? When he handed over the data (and other work) to you, what was the intended outcome? – Pete L. Clark Jun 10 '14 at 3:19
  • 2
    @JackeJR: "On hindsight, we should have discussed authorship right at the start before embarking on anything, planned or unplanned." That's definitely the moral of the story, I agree. If both of you are outside of academia, then you are way outside of my own experience.... – Pete L. Clark Jun 10 '14 at 3:31
  • 4
    I can tell you what I would do, but you have to decide whether it feels at all applicable in your situation: no one should get "locked out" of authorship for their joint work. In the absence of a clear discussion of the nature and requirements of the collaboration, each party who was involved in any way should decide for herself whether to be an author. Ideally each would give an explanation for their decision, and that can be used to decide whether the collaboration should continue. – Pete L. Clark Jun 10 '14 at 3:32
  • 6
    There must be a more poignant way to say "don't start collaborating with someone on a paper if you don't want them to be a co-author on the paper" – Oswald Veblen Jun 10 '14 at 3:33

This doesn't seem to be related much to the fact that the other person is a PI. According to the question, someone collected data, and you have analyzed it.

Unless the data was published elsewhere, you cannot cite it. So you will need to include the person who collected it as an author in order to have a source for the data. It's hard to expect otherwise, if you want to use unpublished data as the source for a paper.

The ICMJE guidelines agree; one of their criteria for authorship is:

Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; AND

The other person has met this point through acquisition of the data.

The most gracious thing to do would be to offer the other person a co-authorship, which will then allow them to meet the other three bullet points in the ICMJE guidelines.

The other person can then decide whether to accept the co-authorship, or whether to take a smaller form of acknowledgment for the data collection.

| improve this answer | |
  • 1
    @JackeJR, but conditions 2 and 3 of the ICMJE guidelines are basically help with the drafting the manuscript and final approval. Many people consider the helping with the drafting aspect to be easily satisfied. – StrongBad Jun 10 '14 at 6:59
  • 2
    Just to point out that this answers you addendum questions: if a dataset has been published, anyone can use it (with appropriate citation). – avid Jun 10 '14 at 8:41
  • 1
    Thanks @StrongBad for the link. That is the point I wanted to make: once someone has satisfied the first point of the ICMJE criteria, it is straightforward for them to satisfy the other two. The person just needs to have a genuine opportunity to revise the content of the paper before it is submitted, and they must approve of its publication. It would seem strange for someone to agree to have a paper published with their name on it, because they collected data, without reading a draft of the paper before it is published. – Oswald Veblen Jun 10 '14 at 11:30

Let me get this straight:

  • Your PI (who is on a different institution than you) made a study and collected some data
  • He gave his data to you
  • He proposed some ways to process the data which did not work out
  • You made a better method and you plan to publish without him.

From what I understand, this is not the way collaboration works. Even the fact that he proposed some methodology to process the data (even if it did not work out) it still proves that this work is a joint work and he did not simply handed the data, in the notion that you can do what you want with it.

After all, those simple facts remain:

  • If you were in his shoes, would you ever collaborate again with a guy who takes your data and does not give you co-authorship, even if you were part of the entire process? The answer is a big NO, because there is nothing to gain from such a collaboration.

  • What do you have to lose by sharing co-authorship? You will be first author, he takes credit and he will willingly share his data again if needed. Then he will be a reference for you which in the long run, means more papers for you.

  • What if your paper gets rejected and he finds out? You got nothing and you just burned a bridge with this man.

  • Imagine the other case. Even, if the other guy was an undergraduate student and he collected the data under your supervision, would not it be unethical not to include this student as a co-author?

Do not be short-sighted. Share co-authorship

| improve this answer | |
  • +1 for the last bullet point. So many times I see on this site undergrads being slighted because they're just interns/undergrads/whatever and they "only" collected data. It's a big deal. – LLlAMnYP Aug 1 '16 at 9:28

I remember seeing a paper or Letter to the Editor discussing the requirements to be an author on the paper. They came up with the 5 types of contributions towards a paper, as I remember they were:

  1. Initial Idea
  2. Data Acquisition
  3. Data Analysis
  4. Write the paper
  5. Revising the paper

As long as someone made a significant contribution toward two of these items they should be an author on the paper.

It completely depends on conventions in your particular area. For example in my field the supervisor is always the last author. I know some fields might have them as first author.

But, what it comes down to is what many others have said... don't burn your bridges. You need collaborations!

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy