On rejection in academia

Finally, after a few years of trying to get work published, I was recently able to get a paper that received the decision, “Acceptable after major revision” for one of the top journals in my field. This was a great relief, as it meant that I could go on the tenure-track job market sooner. However, along the way, I faced a number of rejections. The first paper I wrote as a PhD student was rejected four times before finally getting the decision “Minor revision.” Being slightly more established and experienced now, I have also had the privilege to serve as a peer reviewer for other scholars’ articles. So I have also been on ‘the other side’, so to speak, where I have also sometimes recommended that a paper be rejected. I thought it would be helpful to write a post about my personal experience with rejection, why it happens, and how you can use rejection to your advantage, so that early career scholars and PhD students do not become too discouraged when they are inevitably faced with it.

1. Rejection happens to everyone. Even Nobel Prize-winning scientists and famous scholars get their work routinely rejected. It happens to the best and most talented. The top journals and conferences in most fields have an acceptance rate of less than 20 percent (some under 10 percent), so if you have gotten rejected from one of these premier journals, you are in good company. In most good journals, the acceptance rate will also be sub-30 percent, so rejection is not unusual. In fact, it is the norm. Due to limited journal/conference space and due to the fact that a journal’s reputation is in part determined by its selectivity, Associate Editors are inclined to reject (so a single negative review from a peer reviewer is often grounds for rejection of a manuscript).

2. Rejection does not mean the paper or idea is bad. I used to work in guidance, navigation and control (GNC) engineering, and byfar, one of the most influential ideas used in this area is the Kalman filter. The paper on the Kamlan filter was pivotal in launching the Apollo project, NASA’s first mission to the moon, and it has since been adopted in a variety of other fields like robotics and econometrics. It is used in both industrial and academic settings. Despite its widespread influence, the original paper written by Rudolf E. Kalman was initially rejected.

In my own field of statistics, one of the most influential papers is the paper that introduced the Benjamini-Hochberg procedure on controlling false discovery rate (FDR) in large-scale multiple hypothesis testing. This idea has now been adopted in most fields that apply statistics, but these authors struggled to publish their work initially, facing rejection after rejection. So as you can see, some of the most influential and powerful ideas were initially met with rejection.

3. Often times, rejection can occur because of a mismatch in scope. Rejection may not be the result of a paper being bad, but simply a mismatch in scope. In my field, journals can be generally categorized as:
  • 1) methodological (i.e. proposing new methods that differ from or improve upon existing ones)
  • 2) theoretical (where the underlying theory is the most important innovation)
  • 3) applied (where the specific application to real data sets is the most important aspect).
If you submit a very applied paper to a theoretical journal, then you can expect it to be rejected. Likewise, a paper that contains impressive theoretical results but that is nontrivial to implement in practice or does not advance much methodological work may be ill-suited for a methodological or applied journal.

Additionally, the very top journals will place a high premium on novelty and the potential to have a broad impact on the field. While these are certainly subjective measures, one can usually get a somewhat objective sense of how “novel” something is in relation to existing work. For example, in order to publish in the top methodological journals in my field, the work should greatly improve over existing methods in some way (e.g. it is computationally more efficient, it works well in a greater variety of settings, it remedies a shortcoming with many existing methods, etc.), contain some theoretical justification, and have the ability to be adapted to other problems. For the top theory journals in my field, you typically need to introduce a new theoretical framework, do a theoretical analysis that has never been done before, and/or introduce new tools that can be used to theoretically analyze a broad range of problems.

Some papers may indeed be very good, but they: a) are more narrowly focused, or b) their contribution is considered to be a bit too incremental for the journal in question. For example, if a paper is about a very niche topic within multivariate times series (for example), it may be better suited for more specialized journal like Journal of Time Series Analysis. Similarly, a paper that contains new methodology or theory may be considered to be too “incremental” if it does not advance the state-of-the-art or if it is seen as just a minor improvement over existing methods.

4. Rejection can also occur because of insufficient proofreading and failure to follow directions. If you submit to a journal in my field, you typically need to follow a specific template for journal submission. Failure to follow instructions can result in rejection. Similarly, poorly written manuscripts can also be rejected on these grounds. It is best to have at least one person proofread your work to make sure that the English is as ‘perfect’ as can be before you send it out. Before submitting, you should make sure to follow the journal’s “Instructions for Authors” to a T to avoid rejection for not following directions.

5. Referees and Associate Editors are not out to get you. Typically when you submit to a journal or conference, one of two things will happen. The Editor(s), Associate Editor(s), or conference chair(s) will read the manuscript and may choose to reject the manuscript without sending it out for peer review. You will typically hear back fairly quickly (within a month) if your paper is rejected. At this stage, as many as 50 percent of articles can be filtered out, and the reasons for rejection are often related to points #3 and 4 above.

If the associate editors or conference chairs decide that the paper might be a good fit for publication, they will then send out your paper for peer review. This means that they will contact experts working in the same area as the author(s) to get their opinions on your paper. The decision to reject, ask for a revision, or accept lies solely with the Associate Editor. However, the reviewers can make recommendations as to whether to ask for a revision or to reject the manuscript. After the first round of peer review is completed, the most common decisions are either “Reject (after review)” or “Ask to revise.” Having served as a referee, I can also say that I and most other peer reviewers are not out to get anyone. First, whenever I have agreed to peer review an article, I have never known the authors personally, or I may be only vaguely familiar with them. Given that I don’t know them, I do not have anything personally against them. Second, if I have recommended rejection, it has always been because I found legitimate flaws with the manuscript, which I outline in the next bullet point.

6. The good thing about rejection is that you get a ton of comments. The great thing about rejection is that it usually comes with a long list of comments in the form of a referee report. These comments are meant to help you improve your manuscript, so that you can make the necessary changes and submit an improved paper to another journal. Whenever I have gotten a paper rejected, I may have initially been disappointed. But after reading through the comments, I have usually realized that the concerns raised by the reviewers were actually legitimate. Moreover, after I made changes to my paper, I realized that the new manuscript did markedly improve (sometimes drastically so).

When I have acted as a referee, I have always done my best to provide good feedback to the authors. I always mention some things I liked about the paper and then proceed to list areas for improvement. The rationales for why I have recommended rejection vary, but off the top of my head: sometimes, it has been because I felt that another set of papers had done a very similar analysis and these other papers had been more thorough in their treatment of the problem. Sometimes, it has been because the authors made too many erroneous and technically incorrect claims. Sometimes, it has been because the paper dealt with a problem that has already been addressed thoroughly in the literature, and the authors failed to make a convincing case why their method improved upon existing methods. Whatever the reason, I have always listed both positives and negatives. And of course, depending on the prestige of the journal, I calibrate my expectations, so I will be a bit “tougher” if the journal I’m reviewing for is one of the top four journals in my field, and not so tough if the tier of the journal is a lot lower.

7. There is a bit of noise in the process. Sometimes we just get unlucky with the referees and get an especially harsh reviewer, a reviewer that just does not understand the paper, or a reviewer who rejects a paper for a reason where they themselves made a mistake. It happens. People are human and make mistakes. However, I have found most referees to give good feedback on how to improve a manuscript, even if they recommended rejection or were particularly harsh in their assessment of my work.

I will say also that even if you have a harsh reviewer, they will usually have at least some legitimate criticisms. And if you address their comments adequately, they may even recommend “Ask for revision” if you submit to a different journal and they happen to be the same reviewer on your new manuscript. If you get a reviewer who just didn’t “get” or understand the paper, it is also better to try to make your work as clear as possible. In my work, I like to use bullet points in the introduction to list the specific contributions that my paper makes, so there is no doubt.

Leave Comment

Your email address will not be published. Required fields are marked *