The Art of the Abstract

I’m one of those old-fashioned types who still gets an email from the arXiv every morning notifying me of the latest contributions and listing their abstracts. I still prefer to get my daily update that way than via logging onto the website, although I suspect that’s really force of habit more than anything. The emails are longer these days than they used to be, of course, so now I only manage a quick skim but it’s still a worthwhile exercise.

I have noticed over the twenty-odd years that I’ve been subscribing to this service that as well as being more numerous now, abstracts are also unquestionably longer (at least on astro-ph), to the extent that one sees the dreaded “[abridged]”, indicating that the (approximately 20-line) length limit has been exceeded, much more frequently now than in the past.

Without criticising individual papers, it does seem to me that excessively long and ponderous abstracts are likely to be counter-productive. The whole point of an abstract is that it is a sort of executive summary of the paper which is supposed to convince the reader that the whole paper is worth reading. Given the number of papers there are flying around, a short pithy abstract with a high density of key ideas and results is much more likely to get people reading further than one that waffles on and on about “discussing” and “constraining” this that or the other. Abstracts should be about answering questions, not merely addressing them.

Another mistake that some abstract writers make is to write the abstract as if it were the introduction, which isn’t the point at all. The first few sentences of the abstract should establish why the topic is interesting, but that doesn’t mean it’s meant to be a mini-literature review. References in the abstracts are best avoided altogether, in my opinion.

When so many experienced professional scientists write poor abstracts it’s hardly surprising that our students also struggle to compose good ones for, e.g., project reports. The best advice I can offer is always write the abstract last of all, when you know exactly what is in the rest of the paper. Incidentally, it is often a good idea to write the conclusions first

Once you have finished everything else then set yourself the task of making your abstract as brief as possible but ensure that it answers the following questions (in no more than a couple of sentences each):

  1. Why is the topic of the paper interesting? What is the question you’re answering? Summarize the background.
  2. What did you do? What techniques/data did you use? Summarize the method.
  3. What were your results? Summarize the key results.
  4. What are the wider implications of your results? In particular, how do they answer the questions in 1?

If your abstract comes out more than 20 lines long then cut it. If one of the four sections is much longer than the others then chop it mercilessly to restore the balance. The shorter the abstract the better it is, in my view, although perhaps you don’t have to go this far

Come the revolution, when all papers will be available online, the abstract will be even more important in getting your work recognized. Digital open access publishing will increase the amount of stuff “out there”, and a good abstract is going to be essential to raise your paper’s signal above the noise level.

Abstracts no doubt play different roles in different fields. I understand that in some disciplines abstracts are even actually the primary mode of publication. I think the guidelines above are pretty good for astrophysics, physics generally, and perhaps even most physical sciences. I’d be interested to hear from folk working in other disciplines how they might be modified to suit their requirements, so please feel free to comment below.

Comments will not be abridged.

15 Responses to “The Art of the Abstract”

  1. Anton Garrett Says:

    I very much agree.

    There is a tension between two aspects of an Abstract: hoping to hook the reader into looking at the rest of the paper; and brevity. Science is now so specialised that to cater for physicists in a different part of physics you need to spend a considerable portion of the Abstract doing scene-setting of a sort that specialists in the same field will not need. This lengthens the Abstract. To do or not to do?

  2. Wasn’t there a time when abstracts were not written by the author?

    • Anton Garrett Says:

      As I recall, Einstein used to write abstracts of the work of others early in his career. Lots of these were published together as a service to the scientific community. (This was a century before arXiv.)

  3. telescoper Says:

    Here’s another abstract that’s a little on the short side…

    Click to access GuaranteedRobustnessMargins_for_LQG_Doyle.pdf

  4. One force pushing towards long abstracts is ADS. If you discuss Topic X tangentially in a paper about Topic Y, and Topic X isn’t explicitly in the abstract, then your insight on the connection between Y and X is lost forever because it doesn’t show up in an ADS search of Topic X.

  5. about the very short abstract you link to: i guess that the same abstract is good for *all* the papers which title ends with a question mark

    • telescoper Says:

      You can usually be pretty confident that if a title is in the form of a question, the one thing the paper won’t do is answer it.

      • Actually, if a paper has a question mark in the title, the answer is usually “no”. 🙂 (Exercise for the reader: What about the paper with the title “If a paper has a question mark in the title, is the answer always no?”)

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: