How to Write a Research Abstract: A Worked Example

Check the word limit, required headings and any design-specific reporting guidance. A structured abstract for a trial differs from an unstructured abstract for some qualitative journals.

The ICMJE. Preparing a Manuscript for Submission to a Medical Journal. provides general guidance for preparing medical manuscripts, including abstracts. Use it alongside the target journal’s instructions rather than assuming one structure applies everywhere.

Weak: “This study explored an important public health issue. Quantitative methods were used. Several significant findings were identified, and recommendations are provided.”

Nothing in that paragraph tells a reader what was measured, who participated or what happened. Importance should emerge from the question and evidence, not repeated adjectives.

The following is an entirely fictional teaching abstract, not a real study or publishable evidence.

Objective: To describe seven-day appointment attendance and reported reasons for missed visits among adults attending two outpatient clinics.

Methods: We followed a consecutive sample of 120 adults with scheduled appointments. Attendance within seven days of the scheduled date was assessed using clinic records. Participants without recorded attendance were contacted using an approved follow-up procedure. Analyses were descriptive.

Results: Attendance status was available for 112 participants; 84 attended within seven days (75%). Of 28 participants without recorded attendance, 20 completed follow-up interviews. Twelve cited difficulty leaving work and eight cited transport costs; participants could report more than one reason.

Conclusion: One quarter of participants with known attendance status did not attend within seven days. Reported work and transport constraints identify issues for further investigation, but the study does not establish their causal effects or represent all clinic users.

This example is deliberately modest. It identifies the denominator, distinguishes observation from explanation and does not call descriptive associations causal. A real abstract may require uncertainty estimates and additional design-specific information.

Include the design, setting, participants, important measurements and main analysis. Details that do not affect interpretation can remain in the full paper. However, removing a design feature that changes the meaning of the result is not efficient editing—it is omission.

For qualitative work, identify the approach, relevant participant or data scope, and analysis method in language appropriate to the study. Results should communicate the analytical finding, not just list the topics discussed.

Every major conclusion should be supported by a result stated in the abstract. Do not introduce an intervention recommendation that the study did not evaluate, or claim national relevance from a narrow sample without qualification.

Avoid using “significant” without clarifying whether you mean statistical evidence or practical importance. Where effect estimates are central, report their magnitude and uncertainty rather than only p-values.

Compare the abstract with the final tables and manuscript after every substantial revision. Check sample sizes, denominators, outcome labels, time points, direction of effects and conclusions. Abstracts often retain numbers from an earlier analysis because they were edited first and checked last.

Finally, read only the title and abstract. Can a reader identify what the study did, what it found and what remains uncertain? If yes, the abstract is doing its job. Continue with quantitative results reporting or the publication collection for the full manuscript.