The abstract holds an odd place in a research paper. It’s not really part of the paper itself. It’s more like the elevator pitch. The abstract is usually the last section that you write, but it’s the first part that people will read. Obviously, you want to write an abstract that entices the reader to keep reading. But what if they don’t read beyond the abstract? I recently asked myself this question, and it has fundamentally changed my perspective of abstracts, both as a reader and as an author.
Filtering in Literature Review
Keeping up with current literature can be overwhelming sometimes, and I try to catch any papers that might be relevant. So, in a typical week I might receive alerts for well over 100 papers. I usually start by scanning just the titles, which might eliminate over 95% right away. Then, unless the paper is a definite read (e.g., if it’s clearly in my field, or written by authors whose work I’m already familiar with), then I will read the abstract. If it looks at least somewhat related to my work, then I will add it to “the list” and read it in full (eventually …). If I’m not sold by the abstract, then I move on.
I recently reflected on my filtering process, and came to better appreciate the importance of the abstract. It’s one thing to summarize your paper and make it seem exciting enough for the reader to keep reading. But (and this is the really important part), what if the reader decides to not read your paper? No offense is intended, I’m sure. But did your abstract include the most important information? Will the reader have the best possible perception of your work? Or will they have an impression of your paper that is not what you intended?
Suggestions for Abstracts
With these questions in mind, here are my suggestions for content in abstracts:
- Highlight your main results. A paper shouldn’t be a suspense thriller. Don’t be afraid to give away the ending. If you proved a major result, then go ahead and be up front about it.
- Mention broader applications or generalizations of your work, especially if it extends beyond what is suggested by the title. This kind of statement might be suitable at the end of the abstract. On a related note, the title shouldn’t be so specific that it only applies to part of the paper.
- Favor being explicit over being specific. Abstracts need to be short. You probably won’t have space to include all of your qualifying assumptions. If you have numerical results, then there’s no need to regurgitate specific values. Instead, explicit statements about the general trends are probably most appropriate (e.g., we demonstrate that X performs better than Y in terms of A but at the cost of B).
While all of this might seem obvious, this strategy is something that I think has been missing throughout some of my own work, so it’s an area that I hope to improve on. I’ve left (what I believe to be) important points buried in the middle of manuscripts, or hidden broader applications of my results out of the writing altogether. If I don’t say it myself, then why should I expect you to figure it out? And if you have something important to say, then you might as well be up front about it.