In the Beginning is the End: The Basic Outline of a Scientific Paper, Part 6: Abstract

Tips for Authors:

I know, I know, the abstract comes first, so why am I discussing it last? Because, like the conclusion, it should be based on what you have already written. Indeed, like the conclusion, it should be just rephrased or copied parts from the text, and copying is actually better than rephrasing! With the exception of an abstract for a poster or presentation, abstracts should never be written de novo. However, too often they are, and as an editor, I can tell when the abstract was done without much reference to the actual text.

The abstract should evolve with the text. However, it should not even be attempted until after the first complete draft. I would recommend not even attempting one until you are very close to a penultimate draft or one that you might consider sending to an outside reader. At that point, the paper’s organization should be fairly well set and the terminology of the work clear. Both of these things are important because they need to be reflected in the Abstract.

Unless the journal gives different instructions, an Abstract is roughly one longish paragraph, or 7-9 sentences (this is roughly equivalent to the 30 lines that NIH uses for its standard grant summary). There are two strategies you can use to develop an abstract, once you have a good, complete draft of your paper.

Strategy 1: If you have a Conclusion or concluding paragraph patterned after my suggestion for that section (see my previous blog), then it is just a matter of enlarging upon that. Always, copy the exact wording from the text as much as possible.

1. After the first sentence in the Conclusion, add a little more detail, 1 or 2 sentences’ worth, from the Introduction.

2. Summarize the methodological approach(es) in 2-3 sentences.

3. Expand the Conclusion’s ‘take-home’ message, again with 1 or 2 sentences, to include some of the more interesting points/results behind the overall conclusion.

4. The last sentence from the Conclusion can be the last sentence of the Abstract.

5. Clean up the transitions between sentences and any resulting grammar problems.

Strategy 2: If you do not have a succinct Conclusion, this is what I suggest:

1. Copy 3-5 sentences directly from the Introduction, ones that generalize the research problem, give the general approach you are taking and, especially, the one stating the hypothesis.

2. Summarize the methodological approach(es) used in 2-3 sentences.

3. Copy 4-6 sentences directly from the Discussion, particularly those that generalize the conclusions, are important summaries of the major points/results and give the implications of the research.

4. Consolidate or combine the sentences, making sure that the phraseology is consistent. (This should not be a problem, if you used consistent phraseology in the text!) You should go from 9-14 to 7-9 sentences.

5. Clean up the transitions between sentences and any resulting grammar problems.

At the end you should have an abstract that directly reflects the text.You should have noticed that the resulting structure of the Abstract tracks with the paper: introduction, methods, and results/discussion.Consistency with the text becomes a non-issue because the content of the Abstract comes directly from the text!