Discussing Research Results and Findings
The discussion interprets results, explains their meaning, compares them with previous evidence and identifies implications. It answers “What do these findings mean?” without inventing causes or merely repeating the results.
Learning objectives
- Structure discussion around objectives and principal findings.
- Compare agreement and disagreement with literature.
- Interpret magnitude, precision and clinical significance.
- Explain plausible mechanisms without overstating causality.
- Integrate strengths, limitations and implications.
Discussion structure
- Briefly state the principal finding for the objective.
- Interpret its magnitude, direction and uncertainty.
- Compare with relevant studies.
- Explain plausible similarities or differences.
- Consider bias, confounding, chance and design limitations.
- State implications for practice, policy or research.
Worked paragraph framework
The study found that 39% of patients missed follow-up, indicating a substantial continuity-of-care gap. This estimate is higher than that reported in Study A but similar to Study B. Differences may reflect appointment definitions, urban versus rural access and follow-up periods. Because the design was cross-sectional, associated factors cannot be assumed to cause missed visits. Nevertheless, the pattern supports testing context-sensitive reminder and transport interventions.
Comparison with literature
| Relationship | Discussion task |
|---|---|
| Agreement | Explain shared populations, mechanisms, definitions or settings |
| Disagreement | Examine design, sampling, measurement, period and context |
| Novel finding | Assess plausibility, precision, multiple testing and need for replication |
| Null finding | Consider true absence, low power, measurement error and restricted variation |
Statistical and clinical importance
A small p-value does not establish a clinically important effect. Discuss the effect estimate and confidence interval, plausible benefits and harms, baseline risk, patient priorities, feasibility and cost. A non-significant wide interval may indicate uncertainty rather than no effect.
Causal caution
Association language
“Was associated with,” “was more common among,” or “correlated with.”
Causal language
“Caused,” “led to,” or “resulted in” requires design and evidence supporting causation.
Qualitative discussion
Relate themes to the research question, conceptual framework and existing literature. Explain variations and negative cases, participant context and researcher reflexivity. Do not turn every quotation into a universal conclusion.
Common errors
- Repeating tables instead of interpreting them.
- Citing studies without comparing methods or context.
- Explaining unexpected findings with unsupported speculation.
- Ignoring non-significant or contradictory results.
- Claiming causation from observational association.
- Introducing new results not reported earlier.
- Failing to connect findings to objectives.
Scenario: apparent protective association
A cross-sectional survey finds lower obesity among people reporting frequent exercise. Discussion should consider reverse causation, self-report error, age, illness and diet as possible explanations or confounders. The finding supports an association but not proof that the measured exercise pattern caused lower obesity.
Implications
- Practice: what clinicians or services might reconsider.
- Policy: what systems, standards or resources may require attention.
- Research: what uncertainty needs a stronger design or different population.
- Community: how findings affect participants and equity.
Review questions
- Differentiate results from discussion.
- Discuss a non-significant result with a wide confidence interval.
- Explain four reasons two studies may disagree.
- Rewrite a causal claim from a cross-sectional study.
References
- UAHEB. Research Proposal and Report Writing Guidelines.
- EQUATOR Network reporting guidelines.
Educational content only; interpretation must remain consistent with the study design and evidence quality.
