Most doctoral candidates write a discussion chapter that summarizes findings instead of interpreting them. They restate what the data showed, compare it to a few prior studies, and move on. What they do not do is explain why a finding occurred, how unexpected results change the picture, or what their data means for the theoretical framework on which the study was built. Examiners call this the difference between describing and arguing, and the discussion chapter is where they look most carefully to see which one you are doing.
The evidence for this is specific. A study of 30 experienced doctoral examiners from five universities found that examiner comment was most extensively devoted to the analysis and reporting of findings, the interpretive work at the heart of the discussion, and that examiners "are vigilant for overstated, misrepresented or misreported findings, unconvincing analyzes or interpretations and conclusions that go beyond what the findings can support." The same study found that examiners "expect and want a thesis to pass," but that interpretation is the territory where most fall short. Understanding what interpretation actually requires, structurally and linguistically, is what this guide covers. For support building or reviewing the chapter, our dissertation discussion chapter service works with candidates through both the structure and the argument.
Quick Answer:
A doctoral discussion chapter has eight components: a brief restatement of key findings, interpretation of each finding (the analytical core), comparison with prior literature, connection to the theoretical framework, implications (theoretical, practical, policy), limitations, recommendations for future research, and a conclusion that states the contribution. Interpretation occupies roughly half the chapter and should explain why findings occurred, reconcile contradictions, account for unexpected results, and identify boundary conditions on the claims. Limitations should distinguish deliberate design boundaries from flaws, and claims should be calibrated to evidence using hedging language ("appears to," "suggests," "may"). Examiners want the thesis to pass, but they assess interpretation most closely, and a chapter that summarizes rather than argues does not demonstrate the "doctorateness" the degree requires.
The Difference Between Results and Discussion: Why It Matters
The results chapter reports what you found. The discussion chapter argues what it means. That distinction sounds simple, but most candidates conflate them, either by interpreting findings in the results chapter before presenting them fully, or by summarizing findings again in the discussion when the interpretive work should begin.
The reason the distinction matters is epistemic: results are data; discussion is knowledge. Examiners use the discussion chapter to test whether you can move from observed data to warranted inference, which is the core intellectual act the doctorate is designed to assess. The UK Quality Assurance Agency's Characteristics Statement for Doctoral Degrees frames this as "an original contribution to knowledge in your subject, field or profession." That contribution is made in the discussion, not the results. Everything before it generates the evidence; the discussion makes the argument from it.
In quantitative dissertations, the results and discussion chapters are almost always kept separate, with the objective data reported in the results section and all interpretation reserved for the discussion. In qualitative research, and in some mixed-methods designs, the two may be combined, with each theme's data presented alongside its interpretation. If your program or discipline combines them, the same interpretive requirements apply; the chapter still needs to argue, not just describe.
Table 1: The Eight Components of a Dissertation Discussion Chapter
Component | Purpose | What belongs here | What examiners check |
|---|---|---|---|
1. Restatement of key findings | Orient the reader before interpretation begins | Brief, non-technical summary of the most important findings (2–4 sentences, not a full recap) | Whether this is genuinely brief or whether it duplicates the results chapter |
2. Interpretation of findings | The analytical core converts data into an argument | Explanation of WHY each finding occurred, proposed mechanisms, reconciliation of contradictions, and accounting for unexpected results | Whether interpretation is present at all, and whether it offers a genuine argument rather than a re-description |
3. Comparison with prior literature | Position findings within the existing knowledge base | Agreements, divergences, and explanations for divergences; new literature introduced where findings demand it | Whether contradictions are explained, not just noted, whether the comparison adds to the argument |
4. Relationship to theoretical framework | Close the loop between the study's theoretical premise and its evidence | Whether the data supports, extends, challenges, or complicates the framework, what this implies for the theory | Whether the framework reappears and is engaged substantively, or was mentioned only in the literature review and then abandoned |
5. Implications | State what the findings mean for theory, practice, and policy | Specific, named theoretical implications; concrete, actionable, practical recommendations; policy changes where warranted | Whether implications are specific and arguable, or vague ("this has implications for the field") |
6. Limitations | Accurately scope what the study can and cannot establish | Design boundaries (deliberate scoping decisions) named and explained; genuine flaws acknowledged with mitigation; claims calibrated accordingly | Whether limitations are specific, whether claims stay within them, and whether the candidate understands the difference between a boundary and a flaw |
7. Recommendations for future research | Show how the study opens the research agenda | Two or three specific research directions that follow from the findings and limitations, with named populations, methods, or questions | Whether recommendations are concrete and follow from the work, or are generic placeholders |
8. Conclusion and contribution | State the original contribution to knowledge | A clear statement of what is now known because of this study, why the field should care, and how it connects to the conceptual position of the research | Whether the contribution is specific and non-trivial, and whether it follows from the evidence presented |
The Eight Components of a Discussion Chapter
A doctoral discussion chapter conventionally moves through eight components, and the weight given to each should reflect where the intellectual work lies.
1. Restatement of key findings. Begin with a brief, non-technical restatement of your most important findings, written for a reader who has not just read the results chapter. This should be a paragraph, not a section, with two to four sentences that orient the discussion without duplicating the results. The error to avoid here is restating every finding at the length it was originally reported, which signals that interpretation has not yet begun.
2. Interpretation of each finding. This is the analytical core of the chapter and should occupy roughly half of the total discussion. Interpretation means explaining why a finding occurred, what mechanism or process produced it, not merely noting that it did. It means asking what a result implies about the underlying phenomenon, and whether that implication changes, supports, or complicates the theoretical premise of the study. This is what Trafford and Leshem describe as "intellectualizing, conceptualizing and contributing to existing knowledge" rather than merely reporting. Candidates who describe what they found rather than arguing what it means are, in examiner terms, not yet demonstrating "doctorateness." The literature review you built at the beginning of the dissertation is the standard against which your interpretations are measured; our dissertation literature review support covers how to build that foundation in a form that the discussion can build on.
3. Comparison with prior literature. Connect each interpreted finding to what prior studies found. Point out where your data agrees with the literature, where it diverges, and why the divergence might exist, different population, different context, different method, or a genuine theoretical challenge. The comparison is not a list of who found what; it is an argument about what your findings add to, or revise, the existing picture.
4. Relationship to the theoretical framework. Your study was designed against a theoretical or conceptual framework. The discussion must return to that framework and say what happened to it: does the data support it, extend it, challenge it, or complicate it? A discussion that never mentions the framework again after the literature review has not done the theoretical work the study requires.
5. Implications. State what your findings mean for theory, for practice, and, where relevant, for policy. Theoretical implications say what the findings contribute to how the field understands the phenomenon. Practical implications name specific, actionable recommendations for the people and settings the research addressed. Policy implications specify the legislative or institutional changes the findings could justify. The weakness to avoid is vagueness: "this research has implications for the field" is not an implication; "these findings suggest that practitioners should revise X, because..." is.
6. Limitations. This is the section candidates most often write badly, either by minimizing genuine design constraints or by catastrophizing them to the point of undermining the study. The correct frame is the distinction between a design boundary and a flaw. A design boundary is a deliberate scoping decision, a single site, a cross-sectional design, or a specific population that defines the scope of the claims the study can make. A flaw is an error that threatens the validity of the findings. State both candidly, explain what each means for the scope of your claims, and describe what you did to mitigate the limitations where possible. Do not claim the study has no limitations; that answer tells examiners you have not thought carefully enough about the methodology.
7. Recommendations for future research. Identify two or three specific, concrete directions the study opens for future research. "Future research could explore this topic further" is not a recommendation; "future studies should examine X in Y population using Z method, because this study's cross-sectional design could not establish..." is. Good recommendations follow directly from the study's limitations and findings.
8. Conclusion and contribution. Close the chapter (and the dissertation) with a clear statement of the original contribution to knowledge. This should be a paragraph, not a footnote: what is now known that was not known before, why the field should care, and how the contribution connects to the study's conceptual position. This is the paragraph examiners re-read when assessing whether the doctorate has been earned.
The Interpretive Moves That Distinguish Doctoral Work
Interpretation at the doctoral level is not a matter of adding opinion to data. It is a set of specific intellectual moves, and naming them makes them easier to execute.
Explaining why a finding occurred means proposing a mechanism or process that produced the observed result. "Patients in this group reported lower satisfaction" is a result. "The lower satisfaction scores in this group may reflect the disruption to established care relationships that restructuring produced, consistent with the continuity-of-care literature" is an interpretation. The move is from what to why.
Reconciling contradictory findings means explaining why two things that appear to conflict are both true, or why one is more credible than the other. If your data shows X and the dominant literature shows not-X, you need an argument: is your population different, your measure more sensitive, your theoretical framing more appropriate? A discussion that simply notes the contradiction without explaining it has not done the work.
Accounting for unexpected results means treating surprises as intellectual opportunities rather than embarrassments. An unexpected result is evidence that something in the theoretical model needs revision, or that the context you studied differs from the contexts the literature has described. Explain what the unexpected result implies, not just that it was unexpected.
Identifying boundary conditions means stating when and where the finding holds. All findings generalize to some extent and not to others. Naming the boundary, explicitly and specifically, is a sign of analytical maturity and calibrates your claims to the evidence you actually have. Our dissertation results and analysis support covers how to present the data that grounds these interpretations in a format that the discussion chapter can build on directly.
Table 2: Summary vs Interpretation — The Linguistic Moves That Distinguish Them
Move | Weak version (summary/description) | Stronger version (interpretation) |
|---|---|---|
Reporting a finding | "Participants in the intervention group reported higher satisfaction scores." | "The higher satisfaction scores in the intervention group may reflect the increased continuity of care the protocol provided, consistent with the continuity-of-care literature, which identifies relationship stability as a key driver of patient experience." |
Compared with the literature | "Smith (2021) also found higher satisfaction in structured interventions. This aligns with the current findings." | "The current findings align with Smith (2021), though the effect size in the present study was larger, which may reflect the more intensive contact schedule adopted here — a design difference that warrants direct comparison in future work." |
Handling a contradiction | "Jones (2020) found no significant effect, which contradicts the current findings." | "The discrepancy between the current findings and Jones (2020) is likely attributable to differences in sample composition: Jones recruited from acute care settings where the intervention's relational components are harder to sustain, whereas the present study was conducted in a primary care context. This suggests the protocol's effectiveness may be setting-dependent." |
Stating a limitation | "A limitation of this study is the small sample size." | "This study was conducted at a single primary care site with 42 participants, which limits the transferability of the findings to other settings or populations. Within this context, however, the sample was sufficient to achieve thematic saturation, and the findings are best understood as offering an in-depth account of this specific service context." |
Stating a claim | "This study clearly demonstrates that structured protocols improve patient satisfaction." | "These findings suggest that structured care protocols of the kind implemented here are associated with improved patient satisfaction in primary care contexts, though the mechanisms and their generalizability to other settings warrant further investigation." |
Handling Contradictions, Unexpected Results, and Null Findings
Candidates often approach these situations with anxiety, as though a contradiction or a null result means the study has failed. It does not. It means the study found something, and the discussion chapter is where you say what.
When your findings contradict the existing literature, the discussion should acknowledge the contradiction directly, offer a candidate explanation for it, and evaluate the relative strength of your evidence and the prior evidence. A study with a larger, more diverse sample and a more precisely validated measure may warrant more confidence than earlier work with a narrower scope. Say so, and say why.
When a result was unexpected, treat it as the question it is. Why might this have occurred? What does it imply about the model, the population, the method, or the context? The most intellectually interesting discussion chapters are often the ones where something did not go as predicted, and the candidate engages with it seriously.
When findings are null or non-significant, interpret them rather than apologizing for them. A null finding means the intervention or relationship you examined did not produce the effect predicted, in this context, with this sample. That is information. It can constrain a claim that was previously overstated in the literature, signal a moderating variable, or identify a design sensitivity limitation. Our guide to interpreting and defending null findings covers this in detail and is worth reading before writing this section of the discussion.
Writing Limitations That Examiners Accept
The Golding, Sharmini and Lazarovitch study of examiner practice found that examiners are specifically vigilant for "conclusions that go beyond what the findings can support." That watchfulness is the reason the limitations section exists: it is where you scope your claims accurately and demonstrate that you know what the study can and cannot establish.
The two-part framework for writing limitations is simple. First, name the limitation specifically: "this study used a convenience sample of 47 undergraduates at a single institution" rather than "the sample was limited." Second, explain what it means for the scope of the claims: "the findings are therefore not generalizable beyond this institutional context, though the mechanisms identified may be transferable to similar settings." That formulation is honest, specific, and intellectually defensible. It does not undermine the study; it calibrates it.
A genuine design flaw, such as a measurement error, a data-collection problem, or an unacknowledged confound, should also be stated, because examiners will notice it regardless. Acknowledging it with a clear account of what you did to mitigate it, or what it means for the interpretation, is considerably stronger than leaving it unaddressed.
Working on your discussion chapter, and not sure if your interpretation is doctoral-level? |
|---|
The difference between a discussion that passes and one that gets sent back is usually in the depth of interpretation, not the volume of writing. A specialist can read your chapter, identify where you are describing rather than arguing, and tell you specifically what interpretive moves are missing. Send your discussion chapter for review, and you will have an itemized quote within 2 to 4 business hours, no obligation. |
Hedging and Claim Strength
Every claim in a discussion chapter should be calibrated to the evidence behind it. Over-claiming is one of the specific failure modes the Mullins and Kiley study identifies; examiners penalize conclusions that exceed what the findings can support, and under-claiming undermines the contribution. The calibration device is hedging language.
A hedge ("may," "appears to," "suggests," "is consistent with") signals that a claim is an interpretation rather than a certainty, appropriate to the evidence type and the design. A booster ("clearly," "demonstrates," "proves") signals certainty, which is rarely warranted in social science, health, or education research where findings are context-dependent. The goal is not to pepper every sentence with "possibly" and "perhaps, " which produces the opposite problem: a discussion that fails to commit to any claim, but to match the strength of the linguistic signal to the strength of the evidence. When the evidence robustly supports a claim, state it confidently. When it is directional, preliminary, or context-specific, frame it as such.
Professional Doctorates: How the Discussion Differs
For EdD, DNP, and DBA candidates, the discussion chapter carries an additional weight: it must demonstrate the connection between the study's findings and the practical, organizational, or policy change they warrant. A PhD discussion privileges theoretical contribution; a professional doctorate discussion must show how findings translate into professional action. Implications are not an afterthought in a professional doctorate; they are the central claim that the degree is built to produce. Our guide to MSN vs DNP project differences covers how the degree level shapes what the discussion chapter is expected to produce. Once the discussion is complete, the next step is dissemination: our DNP project dissemination guide covers how to convert the discussion chapter's findings section into a publishable manuscript.
Frequently Asked Questions
What is the difference between the results chapter and the discussion chapter?
The results chapter reports what your data showed: factually, objectively, tied to research questions, without interpretation. The discussion chapter interprets what those findings mean, why they occurred, how they relate to prior literature and theory, what they imply for practice, and what the original contribution is. In some qualitative designs, the two are combined into one chapter, but the interpretive requirements remain the same.
How long should the discussion chapter be?
Length varies by discipline and institution, so check your program's guidelines. As a proportion, the discussion should be substantive, often comparable in length to the results chapter, and in qualitative work, sometimes longer, because the interpretive work is itself part of the method. What matters more than length is that interpretation receives enough space: a two-page discussion of a large dataset is almost certainly insufficient. Before writing the discussion, it also helps to ensure the methodology chapter is in strong shape, since the discussion's interpretive claims must be grounded in the method that generated the data; our guide to why methodology chapters get sent back covers what examiners check there.
Can I introduce new literature in the discussion chapter?
Yes, and you should, if a finding led you to sources that were not central in your literature review. The discussion is the place to bring in literature that contextualizes, challenges, or supports your findings. What you should not do is use the discussion to correct gaps in the literature review that were already apparent from the design stage; examiners will notice, and it signals poor prior planning.
What does "contribution to knowledge" mean in the discussion chapter?
A contribution to knowledge is a specific, non-trivial claim about what is now known or better understood because of your study. It does not have to be a grand theoretical revelation; it can be an empirically supported refinement of an existing model, a demonstration that a finding replicates in a new context, an identification of a moderating variable that prior work missed, or a practical framework that did not previously exist. What it cannot be is a restatement of the findings in different words. The contribution is what your data implies, not what your data showed. If your data produced null or non-significant results, our guide to interpreting and defending null findings covers what the contribution looks like when the expected effect was not found.
What are the most common examiner criticisms of discussion chapters?
Based on empirical studies of examiner reports, the most common criticisms are: interpretation that stays at the level of description rather than argument; implications that are vague or generic; limitations that either over-claim the study's scope or catastrophize minor design constraints; conclusions that exceed what the findings can support; and failure to connect findings back to the theoretical framework. Candidates who anticipate and pre-empt these criticisms in the chapter itself are less likely to encounter them in the examination.
You May Also Find Useful
Argue, Don't Describe
The discussion chapter is the place where the dissertation becomes a scholarly argument rather than a report of activities. Every section should be doing interpretive work: explaining why findings occurred, placing them in conversation with prior literature, connecting them to the theoretical framework, calibrating claims to evidence, and ultimately making the case that the study has produced something worth knowing. Examiners want the thesis to pass. They read the discussion chapter to find out whether you have earned that outcome by demonstrating the capacity to argue from evidence rather than describe it. If you would like a specialist to review your discussion chapter before submission or examination, share what you are working on, and you will have an itemized quote within 2 to 4 business hours, no obligation.
