Working with coauthors

By Geoff Hart

Previously published as: Hart, G. 2018. Working with coauthors. World Translation Services, Japan.

Scientific research is rarely a solitary effort. Most research is now done by teams of researchers, whether from the same university or from institutes around the world. Partly, this is because few researchers have all the skills required to perform a complex multidisciplinary study. Large studies also require more work than a single person can perform. But for most researchers, it’s also more enjoyable and productive to work with colleagues than to work alone.

When you work in a team, the team members become co-authors of any papers that result from the collaborative research. This raises several issues:

Who should write the paper?

The best way to write a paper varies among studies and research groups. Sometimes one author writes the whole paper, particularly for small or simple studies performed primarily by one researcher or for researchers who enjoy writing and are good at it. This may not be possible for large, multidisciplinary, or complex studies. When no one author fully understands every aspect of the research, each author may write the part of the paper that they’re most familiar with. There are many solutions between these extremes. For example, one author may write all of the paper except the statistical analysis section, which will be written by the statistician who designed the study and performed the statistical analysis.

In what order should the names appear?

At some point, you must decide on the order of the author names. This is challenging, because it requires a balance among several criteria:

Each institute has guidelines that may outweigh other considerations. If co-authors work for institutes with conflicting guidelines, the guidelines of the institute that initiated the study or that provided most of the funding usually take precedence.
Funding agencies may have their own requirements that must be followed; for example, the author who received the first grant may need to be the first author.
Many journals ask authors to identify their contributions to a paper, and choose an order based on the importance or magnitude of their relative contributions.
The author who initiated the study, who wrote the paper, or who has the greatest seniority (i.e., has the longest research history) is often chosen as the first author.
When there are many authors, the first three or four may be listed in order of their contributions, but subsequent authors are listed in alphabetical order by family name. This avoids giving offence to those authors by being forced to judge their relative contributions.
Some institutions place the most important author’s name at the end, believing this to be the most prestigious position. Others place the most important name first because this is the name that will forever be associated with the paper. If I am the first author of a multi-author paper, then that paper will be remembered and cited as “Hart et al.”; if I am the last name of six names, few readers will ever remember that I contributed to the paper.

Balancing these criteria requires negotiation and compromises, such as following one institute’s guidelines for the first paper produced by a study and another institute’s guidelines for the next paper, or such as alternating the names of the first two authors.

Who should approve the content?

All authors must be willing to take credit—or accept blame—for published research; any authors who are unwilling should not be listed as co-authors. As a result, journals ask the author responsible for submitting a paper to confirm that all authors have read and approved the submitted version. This is not a simple formality that you can ignore. Authorship has legal and moral consequences, such as the health of patients treated based on the results of medical research and the human and environmental safety of machine, process, and infrastructure designs based on an engineering manuscript.

There are also practical consequences. Because each author has different expertise, no one author will fully understand all aspects of a complex or multidisciplinary study (particularly the parts they did not perform). Research technicians sometimes review their supervisor’s papers because they know details the author does not: the author did not perform the analysis or obtain the measurement and cannot know these details. Each author must rigorously review the parts of the manuscript related to their specialty, but must also carefully read the rest of the manuscript to ensure that it does not contradict, misinterpret, or invalidate other parts of the paper that they were responsible for. The more people who inspect a manuscript, the greater the likelihood of detecting and fixing errors before publication. This is important, because journal reviewers rarely check papers thoroughly; indeed, they cannot, since they did not participate in the experiment.

This is particularly important for the results of numerical calculations, which peer reviewers cannot check without access to the raw data. A calculation error that enters the research literature may mislead all subsequent researchers who base their research (or real-world applications of the research) on the paper. Even if the error is subsequently detected and a correction is published, researchers who only encounter the original article are unlikely to search the literature to find the correction.

An effective approach is for one author to perform the calculations and statistical analyses, and another to independently repeat the calculations and ensure that both results agree. A senior researcher may review the calculations of a junior researcher (e.g., a PhD supervisor reviewing their graduate student’s work); conversely, a junior researcher may perform the second analysis to learn the approach, and will ask for confirmation if they obtain different results from what their supervisor calculated.

Rewards and challenges

There is a final practical reason why all authors must review a paper before submission to a journal: the “curse of knowledge”. When you are intimately familiar with your research, it’s human nature to forget that others will be less familiar with your work. Another person will either lack this curse or have it to a weaker degree, and will therefore be better able to detect implicit assumptions that must be explicitly stated before readers can understand your research. Remember: if your colleagues don’t understand what you wrote, neither will the journal’s reviewers. You can waste a lot of time responding to review comments that can be summarized as “I’m sorry, but I don’t understand”. It's obviously more efficient to eliminate such problems before the peer review.

Working in multi-author research groups can be enjoyable and professionally rewarding, but it leads to challenges you won’t face in single-author research.