Authorship for design and data analysis

Question

I’m a Postdoc. Last year I was given some data from a former postdoc. My PI told me that the data had never been published and. I worked on that project night and day, designed the analysis plan, did tons of analysis, came up with a novel technique (which hasn’t been used in the literature before). I got pretty interesting results and last week when I showed my results to my PI and the former Postdoc, my PI said that she (the former Postdoc) should take the lead of the project and be first author. I’m a bit confused, I didn’t know they had that in mind. The former Postdoc didn’t analyze the data, is it ok she gets first authorship with my design and analysis? She said she could write the manuscript (but of course she’ll need help writing methods because she didn’t do that). I felt bad about this. I do understand that it’s her data from her previous project, but I wish they had told me before I did all the work. Am I wrong, or should I be considered first author?

Answer 1

This sound like a classic case of clinician-statistician interaction. Both sides tend to under-appreciate the contribution of the other, and conflicts like this happen a lot. For the future, the lesson is to discuss authorship as early as possible when engaging on any projects.

For now, I'd add two things to consider:

1. Even assuming that the entire intellectual contribution so far has come from you, the other postdoc would not be entirely freeloading. The first author has a lot of responsibilities, such as deciding on the overall presentation, writing the manuscript, presenting the results in conferences, responding to reviews etc., and I personally think that is a significant intellectual effort. It may be unfair that you were passed over, but in the end the other person will still have to put in a lot of intellectual work, while you are free to move on to other projects.

2. Ask yourself if your analysis method is a sufficiently strong advance to stand as a publication on its own, i.e. could you write a methods paper using some simulated or publicly-available data. If so, pitch that to your PI, along the lines of "I believe I have found some more general methods to improve PET data analysis, and I would like to write that up with a technical focus, in an appropriate journal. By the way, this won't actually need [other postdoc]'s data". The PI will surely see the cause of this, but as long you're being honest and tactful, nobody should be offended, and whatever the outcome, this should help all parties understand each other's contribution better.

Answer 2

Let me formalize my comment. The effort someone puts in to a paper may have a lot or only a little to do with the intellectual content. In the worst case, the work is just grunt work: watching solutions boil and bubble and taking notes.

The more important questions for lead authorship is who drove the intellectual content here. For example, who first noticed that this is an important and "interesting" set of questions to ask? Who formulated the hypotheses? Who designed the experimental structure if it wasn't completely standardized. Who defined the standard of "success" in answering the important questions.

The person(s) who do that are the authors. Designing programs to analyze the data may be as important as that, or not, depending on the needed innovation.

I'm not going to try to judge your case (and couldn't, in any case), but those are the important questions.

I hope it is something more essential than lab politics that is the determinant here, but that happens also.