The discussion answers the question you left open in the introduction: what do your results mean? It is the place to interpret, compare and acknowledge limits (Jawaid & Jawaid, 2019). The conclusions close with the essentials.
What it should contain
- Main findingWas the hypothesis supported? A summary of the central result.
- ComparisonAgreements and differences with other studies, and their theoretical explanation.
- Unexpected findingsPossible explanations, without forcing them.
- ImplicationsFor theory and practice: who can use these results and how.
- LimitationsWhat restricts the conclusions and how future studies can overcome it.
- ConclusionsA direct answer to the aim, with no new data.
How it should sound
- Interpretive, not repetitive. Don’t list the numbers again: explain what they mean.
- In dialogue with the literature. Every important finding is compared with other studies.
- Careful with causality. In non-experimental designs, “was associated with” or “predicted”, never “caused”.
- Honest about limits. Real limitations, with the direction in which they could bias the results. In a sample of articles from high-impact journals, only 17% mentioned their limitations (Ioannidis, 2007).
Before and after
| Weak version | Problem | Improved version |
|---|---|---|
| Social media use lowers self-esteem. | Causality in a cross-sectional study | Heavy social media use was associated with lower self-esteem. |
| The only limitation was the sample size. | Generic limitation, with no consequences or remedy | Convenience sampling at a single university limits generalization; future studies could use probability samples from several institutions. |
| In conclusion, 62% of participants showed… | New data in the conclusions | In conclusion, academic stress is related to performance in this population, which supports… |
What each journal decides. Whether limitations get their own subheading, whether conclusions are a separate section or the last paragraph of the discussion, and the length of each part change from one journal or conference to another. Check their author guidelines.
Before you call it done
- Does it begin by saying whether the aim or hypothesis was supported?
- Does it compare the main findings with other studies?
- Is its language consistent with the design, without claiming causality?
- Does it present concrete limitations and how to overcome them?
- Do the conclusions answer the aim without new data or citations?
The Manuscript Diagnosis flags causal language in non-experimental designs, generalizations and incomplete limitations.
References
Ioannidis, J. P. A. (2007). Limitations are not properly acknowledged in the scientific literature. Journal of Clinical Epidemiology, 60(4), 324–329. https://doi.org/10.1016/j.jclinepi.2006.09.011
Jawaid, S. A., & Jawaid, M. (2019). How to write introduction and discussion. Saudi Journal of Anaesthesia, 13(Suppl. 1), S18–S19. https://doi.org/10.4103/sja.SJA_584_18
von Elm, E., Altman, D. G., Egger, M., Pocock, S. J., Gøtzsche, P. C., & Vandenbroucke, J. P. (2007). The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. PLoS Medicine, 4(10), e296. https://doi.org/10.1371/journal.pmed.0040296