Abstracting the Truth


The most read words we'll ever write are our abstracts. Hopefully for years to come, final-year undergraduates all around the globe will be including your work in their reports based entirely on those short paragraphs. I'd like to say that, during my first degree days, I never once committed the cardinal sin of citing a paper that I hadn't actually read, but, well, um, if the journal's not in the library, what can you do?

Once or twice during my fleeting but happy days as an undergraduate, we were passed a paper with the abstract blotted out, the object being for us to have a go at writing one. At the time I remember finding this task immensely difficult. I must have stared at the blank computer screen for an age before I stumbled on the way to attack these puzzles.

The trick for me was learning that most of the important stuff in papers goes into the first and last sentences of each paragraph or section. I know this must be true, as I read it in a cheap paperback book on how to bluff your way with speed-reading. Armed with this knowledge, I soon extracted the juice from the "beheaded" paper by topping and tailing each section. Then I simply rewrote bits of the selected sentences (to get around that plagiarism thing), and the abstract simply self-assembled before my eyes. OK, OK, so my tutor said it was a bit stilted to read, but he had to admit that I hadn't missed a single salient point.

Three years later I was charged with a different and more onerous task: to write the abstract to my first first-author paper. Not knowing where to start I tried the same technique, but I was no longer happy to write anything that might be called stilted. So I concentrated on writing the thing in one smooth, flowing movement and realised that for this you need to know your work inside out, which explains why I couldn't do it as an undergraduate.

Now it flowed, but I was no longer confident I'd included everything. To get around this, I scanned through the results section to pick out the highlights, cross-checking the abstract to make sure I hadn't undersold my story. And suddenly there it was: all the salient points linked in a flowing style. It's amazing how presenting your paper in a nutshell makes the real worth of your discovery shine.

Perhaps the ultimate condensation challenge is writing your thesis abstract. I found the sheer quantity of information I needed to summarise into 200 words rather daunting. As the only criterion you can use to assess what to include is relative importance, you have to accept that 6 months hard graft may have to be turned into four words, or that some results are to be lost forever in the depths of chapter five.

Writing an abstract for a conference is a different matter altogether. Like a poster, this sort of abstract is an advert for your work, not a peer-reviewed piece. Despite your secret desire to put people off coming to your talk lest they ask awkward questions, you want to attract a big audience. On the other hand you don't want to make your abstract so juicy that you either give away all your secrets in advance or commit to something you're not sure about. This is especially true if you have to submit your abstract months before the meeting. I find the best way around this is to leave your readers hanging at the end of the abstract waiting for the talk to fill in the punch line.

Whatever the type of abstract you have to tackle, the following simple structure should provide a good starting point:

  • Write a sentence or two briefly presenting the system you worked on and possibly the essence of the knowledge to date. You can't summarise the entire introduction; you just need to encapsulate the state of play before your paper came along. Make it absolutely clear what you did and what other people did, for example by using phrases such as "It has previously been shown that ..." and "Here we tested ...".


  • Summarise, in a sequence broadly similar to that of your Results section, what your findings are, making sure you refer to the techniques you used. It's the difference between writing "We further showed that X is dependent on Y" and writing "Paramagnetic bipolar tolography revealed that X was dependent on Y." This section should form the bulk of the abstract, so cram as many key words as possible in there, as the five or so key words that journals allow you to specify as such will soon get used up.

  • State in the most succinct and descriptive terms exactly what your results mean: "These results provide the first evidence that ..." If you and your co-authors really believe you have shown something new, then be bold and claim it. Far more people will read this claim than will ever assess your evidence for it by reading the full text of your paper.

  • Follow Science Careers

    Search Jobs

    Enter keywords, locations or job types to start searching for your new science career.

    Top articles in Careers