Writing an effective abstract

Nonlinear Dynamics Laboratory

Writing an effective abstract

The abstract is the part of a paper most people read, and often the only part. Editors read it first when they decide whether to send a manuscript for review. Write it last, when you know what the paper shows.

What an abstract contains

One paragraph, roughly one or two sentences per item:

  1. Context. The problem and why it matters.
  2. Gap. What current methods cannot do.
  3. Approach. What you did: the system, the method, and whether the work is experimental, numerical or analytical.
  4. Results. The main findings, with numbers where you have them.
  5. Significance. What the results change, and for whom.

Most of the words belong to the approach and the results. A reader should be able to state your main finding without opening the paper.

Rules

  • State findings, not topics. Write “the backbone curve stiffens by 12% over the tested range”, not “the backbone curve is investigated”.
  • No citations, figure references or undefined abbreviations.
  • Equations only if the paper cannot be described without one.
  • Stay within the journal’s word limit, and check whether it wants a structured abstract.
  • Reuse the paper’s key terms so that search engines and indexing services find it. Choose keywords that are not already in the title.
  • Check every claim in the abstract against the paper. Nothing may appear in the abstract that the paper does not show.

Highlights and graphical abstracts

Some journals also ask for highlights or a graphical abstract. Elsevier journals, for example, ask for three to five short bullet points with a strict character limit. Follow the journal’s guide for authors. Write highlights as findings, not as a table of contents, and build a graphical abstract from one key figure in the NLD style.

Further reading