How to Write the Discussion Section of a Research Paper: Structure, Outline, Examples

Key Takeaways

• The Discussion section interprets your findings rather than repeating them; every sentence should answer the question, “So what does this mean?”
• Acknowledge limitations honestly: reviewers will identify them regardless, and proactive disclosure strengthens your credibility.
• The structure of a Discussion varies meaningfully by methodology; qualitative, quantitative, mixed methods, and literature review papers each follow distinct conventions.
• A conference paper Discussion follows the same logic as a journal article Discussion but must be compressed to fit strict word or page limits, often prioritizing implications over extensive limitations.

What Is the Discussion Section and Why Does It Matter?

The Discussion section is where you explain what your findings mean. It transforms raw results into knowledge by connecting them to prior literature, your research questions, and the broader field. Without a strong Discussion, even technically sound results fail to persuade.

The Discussion section serves four core purposes:

• Interpreting findings: explaining what the results mean, not merely restating what they are.
• Contextualizing findings: comparing results to prior research and theoretical frameworks.
• Acknowledging limitations: demonstrating scholarly integrity through honest appraisal.
• Stating implications and recommendations: answering the question, “Who should care, and why?”

How Do Results and Discussion Sections Differ?

The Results section reports what happened; the Discussion section explains what it means. The table below contrasts how the same finding is handled in each section.

Dimension	Results Section	Discussion Section
Purpose	Report findings objectively.	Interpret and explain findings.
Voice	Neutral, descriptive, past tense.	Analytical, argumentative, present tense for established facts.
Statistical detail	Full statistics: means, SDs, confidence intervals, p-values, effect sizes.	Cite key statistics briefly to anchor interpretation; do not repeat all figures.
Length of treatment	Each finding gets equal treatment regardless of importance.	More important or surprising findings receive more interpretive space.
References to literature	Rarely cited; data speak for themselves.	Frequently cited; findings are compared to prior studies throughout.
Example: same finding (Group A scored 15 points higher than Group B, p < .01, d = 0.72)	“Group A (M = 82.4, SD = 6.1) scored significantly higher than Group B (M = 67.3, SD = 7.4), t(98) = 10.2, p < .01, d = 0.72.”	“The large effect size (d = 0.72) suggests the intervention produced a practically meaningful improvement, consistent with similar programs reported by Jones et al. (2021), though larger than the moderate effects found by Smith (2020).”
New information permitted?	No: only report what was measured.	No new data; new interpretations and connections to literature are expected.
Tense	Past tense throughout.	Past tense for your study; present tense for established knowledge.

Recommended Structure for a Discussion Section

Portion	Function	Typical Length
1. Opening statement	Restate the research problem and briefly summarize the most important finding. Do not repeat all results.	1 short paragraph
2. Interpretation of findings	Explain what each key finding means. Move from most to least important, or follow the same order as the Results section.	Largest portion of the section
3. Comparison to prior literature	Show how findings agree with, extend, or contradict existing research. Cite relevant sources.	Woven throughout Move 2
4. Acknowledgment of limitations	Identify constraints on validity, scope, or generalizability. Frame each as context rather than failure.	1 to 2 paragraphs
5. Implications	Explain the practical, theoretical, or policy relevance of the findings for identified audiences.	1 paragraph
6. Recommendations	Propose concrete next steps for practice, policy, or future research.	1 paragraph or bulleted list
7. Closing statement	End with a strong synthesis sentence that affirms the contribution of the research.	2 to 3 sentences

Writing the Discussion Section for a Quantitative Study

• Distinguish statistical significance from practical significance: a p-value tells you the result is unlikely to be chance; effect size tells you how large the difference actually is. Discuss both.
• Explain unexpected findings: do not ignore results that contradict your hypotheses. These often generate the most valuable insights and signal integrity.
• Address confounds: acknowledge variables not controlled for and explain how they might have influenced the results.
• Avoid over-generalizing: be precise about the population to which your findings apply.

Checklist: Quantitative Discussion

Checklist Item	Done?
Restated the primary research question and direction of main finding
Interpreted the practical significance of the effect size, not only the p-value
Compared findings to at least two prior studies; noted agreement or contradiction
Explained unexpected or null results rather than minimizing them
Identified at least two specific methodological limitations (e.g., sample characteristics, measurement tools)
Stated what the findings imply for theory or practice
Made at least one concrete recommendation for future research
Avoided introducing new data not reported in the Results section

Taking the help of a professional academic editor, e.g., through Editage’s Premium Editing service, can help you get a second, unbiased opinion on your Discussion and whether it’s thorough enough.

How Should Null or Negative Results Be Discussed?

Null results, those in which no significant difference or relationship was detected, are valid and important contributions to knowledge. Discuss them fully. First, rule out methodological explanations such as insufficient power or unreliable instruments. Then consider whether the null finding itself contradicts or refines prior theory. Never treat a null result as a failed study.

Sample Phrases: Quantitative Discussion

Function	Sample Phrase
Opening interpretation	“The findings support the hypothesis that [X], with participants in the intervention group demonstrating significantly higher [outcome] than controls.”
Comparison to prior research (agreement)	“This result aligns with the work of [Author, Year], who similarly found that [related finding] in a comparable population.”
Comparison to prior research (disagreement)	“In contrast to [Author, Year], the present study found no significant relationship between [X] and [Y], possibly due to differences in [contextual factor].”
Interpreting effect size	“Although statistically significant, the small effect size (d = 0.21) suggests that the practical magnitude of this difference may be limited in applied settings.”
Addressing null results	“The absence of a significant effect for [variable] may reflect insufficient statistical power given the sample size, rather than a true null relationship.”
Noting a confound	“Results should be interpreted with caution, as [variable] was not controlled and may have independently influenced [outcome].”

Writing the Discussion Section for a Qualitative Study

Qualitative Discussions interpret themes, patterns, and meanings rather than statistics. The emphasis shifts from significance testing to richness of interpretation, reflexivity, and contextualization within participants’ lived experiences.

Core Priorities for Qualitative Discussions

• Foreground participant voice: reference representative quotes or vignettes to anchor interpretations, even though the quotes themselves belong in the Results section.
• Engage with reflexivity: explain how your positionality as a researcher may have shaped interpretation. Readers of qualitative work expect this.
• Discuss transferability rather than generalizability: explain the conditions under which findings might apply elsewhere, without claiming universal applicability.
• Connect themes to theoretical frameworks: show how your themes confirm, challenge, or extend the conceptual literature.
• Acknowledge disconfirming evidence: discuss themes or cases that did not fit the dominant pattern.

Checklist: Qualitative Discussion

Checklist Item	Done?
Interpreted the meaning and significance of each major theme
Connected themes to the theoretical or conceptual framework
Compared findings to prior qualitative and quantitative literature
Addressed reflexivity and researcher positionality
Discussed transferability with specific contextual conditions
Acknowledged disconfirming cases or minority perspectives
Addressed trustworthiness: credibility, dependability, confirmability, and transferability
Stated implications for practice or policy relevant to the study context
Made recommendations for further research, including suggested designs or populations

Editage’s Premium Editing service can help you adjust your phrasing for qualitative results so that you’re neither overstating nor unnecessarily downplaying them.

Sample Phrases: Qualitative Discussion

Function	Sample Phrase
Opening interpretation	“The overarching theme of [theme name] reflects participants’ shared experience of [phenomenon], which has been underexplored in prior qualitative literature.”
Connecting to theory	“These findings extend [Author’s Year] theoretical framework by illustrating how [concept] operates differently in [context].”
Addressing reflexivity	“As a researcher with [relevant background], I acknowledge that my perspective may have sensitized me to [particular interpretation]; member checking was used to verify the accuracy of emerging themes.”
Transferability statement	“While these findings are not intended to be generalized, they may hold relevance for [similar contexts] where [key contextual conditions] are present.”
Disconfirming case	“Notably, two participants described experiences that diverged from the dominant theme, suggesting that [moderating factor] may alter the experience of [phenomenon] for some individuals.”
Implication for practice	“The centrality of [theme] to participants’ decision-making suggests that practitioners should prioritize [action] when working with [population].”

Writing the Discussion Section for a Mixed Methods Study

A mixed methods Discussion must do more than report qualitative and quantitative findings separately; it must demonstrate integration, showing how the two strands of data illuminate, corroborate, or complicate each other. This integration is the defining contribution of mixed methods research.

Integration Strategies

Strategy	When to Use It	Example Statement
Convergence	Qualitative and quantitative findings point in the same direction.	“Survey data indicating that 78% of participants reported improved confidence were corroborated by interview findings, in which participants consistently described feeling more capable after the program.”
Divergence	Findings from each strand conflict or differ.	“Although quantitative results showed no significant improvement in [outcome], qualitative data revealed that participants perceived meaningful personal change, suggesting that the survey measure may not have captured [dimension].”
Expansion	Qualitative findings explain or elaborate quantitative patterns.	“The quantitative finding that participants in rural areas showed lower engagement was illuminated by interview data, which surfaced transportation and scheduling barriers not captured in the survey.”

Checklist: Mixed Methods Discussion

Checklist Item	Done?
Summarized key findings from each strand separately before integrating them
Explicitly stated whether findings converge, diverge, or expand each other
Explained what the integration reveals that neither strand alone could show
Discussed limitations of each methodological strand separately and in combination
Addressed how quality criteria for both quantitative and qualitative components are met
Stated implications that draw on the integrated findings rather than treating strands in isolation
Recommended future studies that might further integrate or refine findings

Sample Phrases: Mixed Methods Discussion

Function	Sample Phrase
Convergence	“The convergence of quantitative and qualitative findings strengthens confidence in the conclusion that [finding], as both strands independently pointed toward [interpretation].”
Divergence	“The divergence between statistical outcomes and participant narratives suggests that [construct] may be multidimensional, with different facets captured by each data type.”
Expansion	“Qualitative data provided explanatory depth for the quantitative pattern, revealing that [mechanism or process] accounts for the observed difference.”
Integration statement	“Taken together, these findings offer a more complete picture than either method alone could provide, revealing both the extent of [phenomenon] and the mechanisms through which it operates.”

Writing the Discussion Section for a Literature Review or Systematic Review

A standalone literature review or systematic review replaces a traditional Results section with a synthesis of the existing evidence. The Discussion is therefore less about interpreting primary data and more about evaluating the body of literature, identifying gaps, and drawing conclusions about the state of knowledge.

Core Priorities for Literature Review Discussions

• Synthesize rather than summarize: identify patterns, contradictions, and trends across the body of literature rather than describing each study sequentially.
• Evaluate quality of evidence: comment on methodological strengths and weaknesses across included studies.
• Identify gaps: explicitly state what the literature has not yet addressed.
• State conclusions carefully: avoid claiming certainty beyond what the evidence supports.

See also: Discussion vs Conclusion in a Research Paper

Checklist: Literature Review Discussion

Checklist Item	Done?
Synthesized the main body of evidence across studies rather than summarizing each study individually
Identified points of convergence across the literature
Identified contradictions or inconsistencies and offered possible explanations
Evaluated the overall methodological quality of included studies
Stated clearly what gaps in the literature remain
Noted limitations of the review itself (e.g., search scope, publication bias, language restrictions)
Drew conclusions proportionate to the strength of the evidence
Recommended directions for primary research to fill identified gaps

 

Sample Phrases: Literature Review Discussion

Function	Sample Phrase
Synthesis of convergent evidence	“Across the majority of included studies, [finding] emerged as a consistent pattern, suggesting that [conclusion] is well-supported in the literature.”
Acknowledging contradictions	“Several studies reported [finding A], while others found [finding B]; this inconsistency may reflect differences in [population, measurement approach, or context].”
Evaluating quality	“The overall quality of evidence in this area is moderate; most included studies relied on cross-sectional designs, limiting causal inference.”
Identifying a gap	“Notably, no studies to date have examined [topic] in [population or context], representing a meaningful gap given the prevalence of [issue].”
Limitation of the review	“This review is limited to English-language publications, which may introduce publication bias and exclude relevant findings from non-Anglophone research contexts.”
Concluding with proportionality	“Based on the available evidence, it appears likely that [conclusion]; however, stronger causal claims await replication in experimental designs.”

How to Discuss Limitations Effectively

Limitations should be discussed honestly and specifically, not vaguely or defensively. The goal is to help readers correctly interpret and apply your findings, not to undermine confidence in your work. Frame every limitation in terms of its actual impact on the findings.

Common Categories of Limitations

Category	Examples	How to Frame It
Sample	Small size, convenience sample, single site, limited demographic diversity	“The sample was drawn from [population and location], which limits the applicability of findings to [other populations].”
Measurement	Self-report bias, unvalidated instruments, single time-point data collection	“Reliance on self-reported data introduces the possibility of social desirability bias, which may have inflated scores on [measure].”
Design	Cross-sectional design, lack of control group, absence of randomization	“The cross-sectional design precludes causal inference; longitudinal studies are needed to establish whether [X] precedes [Y].”
Context	Study conducted in a specific policy environment, time period, or cultural context	“These findings were obtained during [period or context], and may not replicate in different policy or organizational environments.”
Researcher	Positionality, potential for bias in qualitative interpretation (qualitative studies)	“My professional background in [field] may have sensitized me to [themes]; reflexive journaling and peer debriefing were used to mitigate this influence.”

Writing About Implications

Implications explain who should care about your findings and why. Separate implications into categories relevant to your field: theoretical, practical, policy-related, and methodological. Not every study will have implications in all categories.

Types of Implications

Type	Definition	Example
Theoretical	Advances, challenges, or refines existing theory or conceptual frameworks.	“These findings challenge [Author’s] model by suggesting that [variable] operates as a moderator rather than a mediator of [relationship].”
Practical	Provides actionable guidance for practitioners, organizations, or individuals.	“Clinicians working with [population] should consider [approach] in light of the finding that [result].”
Policy	Has implications for institutional, governmental, or organizational policy.	“Policymakers seeking to improve [outcome] may consider directing resources toward [mechanism], given the strong association found in this study.”
Methodological	Demonstrates a new approach or reveals limitations of existing research methods.	“The divergence between self-reported and behavioral outcomes suggests that future studies should employ multiple measurement modalities to capture [construct].”

Writing Recommendations

Recommendations are concrete, actionable proposals that follow from your findings and implications. They differ from implications in being directive rather than descriptive. Distinguish between recommendations for practice and recommendations for future research.

The Discussion Section of a Conference Paper

Conference papers follow the same logical structure as journal articles but must compress the Discussion substantially to meet word or page limits, typically 4 to 8 pages total. Prioritize implications and conclusions over exhaustive limitations lists.

How Does a Conference Paper Discussion Differ From a Journal Article Discussion?

The key differences are length, depth, and audience purpose. Conference papers introduce findings to a live audience and invite dialogue; they do not need to stand fully alone in the way a journal article does. The table below summarizes the key contrasts.

Feature	Journal Article Discussion	Conference Paper Discussion
Length	800 to 1,500 words (median ~1,115 words)	200 to 500 words; sometimes combined with Results
Depth of interpretation	Each major finding interpreted in detail	Only the most significant findings interpreted
Literature comparison	Multiple prior studies compared throughout	1 to 2 most relevant comparisons
Limitations	At least 2 to 4 specific limitations discussed	1 to 2 key limitations acknowledged briefly
Implications	Multiple categories discussed in detail	One or two most relevant implications emphasized
Audience purpose	Convinces a skeptical peer reviewer	Engages a live audience; prompts questions
Combined sections?	Results and Discussion always separate	Sometimes merged as “Results and Discussion”

Tips for Writing a Conference Paper Discussion

• Lead with the single most important or novel finding; this is what your audience will remember.
• Use one paragraph per major point: interpretation, comparison to prior work, and implications.
• Acknowledge limitations in a single sentence or two; do not dedicate a full paragraph to them.
• End with a strong, memorable statement of contribution: what does your work add to the field?
• If space permits, include one future research direction that is directly tied to a limitation.

Language and Style in the Discussion Section

Tense Conventions

What You Are Referring To	Recommended Tense	Example
Your own study’s findings	Past tense	“Participants in Group A reported higher levels of [X].”
Established scientific knowledge	Present tense	“Research consistently shows that [X] is associated with [Y].”
Prior studies	Past tense	“Jones et al. (2021) found that [X].”
Implications and recommendations	Present or future tense	“These findings suggest that [action] is warranted.” / “Future research should examine [topic].”

Hedging Language

Hedging language protects against overclaiming while still communicating significance. Match the strength of your language to the strength of your evidence.

Level of Certainty	Sample Hedge Phrases
Strong (well-replicated evidence)	“These findings demonstrate that… / The results confirm that… / The data clearly indicate that…”
Moderate (suggestive but not definitive)	“The findings suggest that… / Results are consistent with… / The data point toward…”
Tentative (exploratory or limited evidence)	“It appears that… / It may be that… / Preliminary evidence suggests… / One possible explanation is…”
Speculative (hypothesis for future testing)	“Future research might explore whether… / It is plausible that… / This pattern warrants further investigation.”

Common Mistakes to Avoid

Mistake	Why It Is Problematic	How to Fix It
Repeating the Results section	Readers already know what the data showed; restating it wastes space and signals a lack of analytical depth.	For each finding, write one sentence anchoring it, then move immediately to interpretation and comparison.
Introducing new data	The Discussion is for interpretation, not new evidence; adding new results confuses the structure.	Return new results to the Results section; if they are supplementary, move them to an appendix.
Overclaiming significance	Assertions beyond what the data support damage credibility and may draw peer reviewer critique.	Use hedging language calibrated to the strength of your evidence; cite effect sizes alongside p-values.
Generic limitations	Statements like “the sample was small” without context offer no useful information.	Specify the sample size, why it is small relative to the research question, and what the impact is.
Ignoring contradictory prior work	Failing to address studies that found different results makes the Discussion appear cherry-picked.	Acknowledge contradictory findings and offer a plausible explanation for the discrepancy.
Omitting implications	A Discussion without implications leaves readers asking, “So what?”	Dedicate at least one paragraph to what the findings mean for practice, theory, or policy.
Losing structure and logical flow	Jumping between unrelated ideas creates reader confusion and undermines the argument.	Use topic sentences, transition phrases, and a consistent sequence (finding, meaning, comparison, so what).

See also: How to Write an IMRAD Paper: Outlines, Examples, Checklists

Frequently Asked Questions

How long should the Discussion section be?

Based on large-scale data from tens of thousands of published articles, the median Discussion section is approximately 1,115 words, or around 7 paragraphs. Most published Discussions fall between 824 and 1,480 words. The Discussion typically accounts for around 30% of the total word count of a research article, roughly equivalent in length to the Results or Methods sections. However, always defer to the target journal’s author guidelines, as word limits vary considerably.

Can I use the first person in the Discussion section?

Yes, in most disciplines, use of the first person (“We found…” or “I argue…”) is accepted and increasingly encouraged in qualitative and mixed methods research and across many STEM fields. Some journals in highly conservative disciplines such as certain areas of medicine or law still prefer passive constructions. Always check the target journal’s style guide. Where first person is permitted, it often produces clearer, more direct prose and is preferred by many editors.

Is it acceptable to include tables or figures in the Discussion section?

Generally, the Discussion does not include new data tables or results figures; those belong in the Results section. However, a summary table that synthesizes your findings in relation to prior literature, or a conceptual framework diagram that you have developed based on your interpretations, can be appropriate in the Discussion. Always check journal formatting guidelines, as some explicitly restrict figures in the Discussion.

Should recommendations for future research go in the Discussion or the Conclusion?

Either location is appropriate, and conventions vary by discipline and journal. The important rule is consistency: do not repeat recommendations verbatim in both sections. If your Discussion already addresses implications and recommendations in depth, the Conclusion can simply point back to them briefly. If your Conclusion is where you synthesize the overall contribution of the study, placing the most important future research recommendation there gives it visibility. When in doubt, follow the conventions of highly cited papers in your target journal.

My Discussion keeps turning into a summary of the Results section: how do I break the habit?

This is the most common problem novice researchers face. Try this diagnostic: read each sentence in your Discussion and ask whether it could appear unchanged in the Results section. If the answer is yes, it is a reporting sentence rather than an interpretive one. For every finding you mention, force yourself to follow it immediately with one of three moves: (1) explain why this finding occurred, (2) compare it to what prior research found, or (3) state what it means for theory, practice, or the next study. Those three moves are the Discussion; everything else is the Results. Editage’s Premium Editing service is designed to assist authors with this issue; an expert editor can tell you which results have been repeated in too much detail in the Discussion section and whether you’ve sufficiently interpreted/expanded on your results or merely restated them.

How do I handle findings that contradict what I expected or what prior literature predicted?

Unexpected findings are often your most valuable findings and should be discussed with care and intellectual curiosity rather than minimized or explained away. First, rule out methodological causes: could the result reflect a measurement issue, a sample characteristic, or a procedural factor? If not, consider whether the finding challenges a theoretical assumption, applies to a context where prior findings may not generalize, or reveals a moderating variable not previously considered. End with a specific hypothesis or research design that could help resolve the contradiction in future studies.