The discussion chapter is where a dissertation earns its argument. This guide is written for PhD and DNS nursing students who have results in hand and need to interpret them, connect them to the existing literature, and state limitations and implications honestly, without overclaiming what the study can support.
Quick answer. A discussion chapter interprets what the results mean, situates them against prior literature, and argues for their significance. It does not repeat the results or introduce new findings.
Students who struggle with the discussion chapter usually struggle because they are still writing results. The two chapters have different jobs, and confusing them is the single most common structural problem committees flag at this stage.
The results chapter reports what was found, factually and without interpretation. It presents the data, the statistical output or the qualitative themes, and states them plainly: this variable was associated with that outcome, this theme appeared in this many interviews, this test reached or did not reach significance. It does not explain why, and it does not say what the finding means for practice, theory, or future research. If you have not yet written that chapter, our dissertation results chapter guide covers how to report findings cleanly before you get to this stage.
The discussion chapter takes those same findings and does the interpretive work. It asks what the finding means, how it fits with what other researchers have found, why it might have turned out the way it did, and what someone reading it should do differently as a result. If a sentence in your discussion chapter could have been written before you collected any data, it belongs in the introduction or literature review, not here. If a sentence introduces a number or a result that was never mentioned in the results chapter, it is misplaced and needs to move back.
Read each paragraph in your draft discussion chapter and ask two questions: does it report a new number or theme not already in chapter four, and does it explain, compare, or extend rather than restate? A paragraph that fails the first test needs to move to the results chapter. A paragraph that fails the second test needs to be rewritten so it does interpretive work rather than repeating a sentence you already wrote.
Programs vary in how they label sections, and you should follow your own handbook first. The structure below is a common, defensible pattern that covers what most committees expect to see, in a sequence that builds logically from the study to its wider significance.
| Section | Purpose | What goes wrong when it is skipped |
|---|---|---|
| Restatement of purpose and key findings | Reorient the reader briefly before interpreting | Reader has to flip back to chapter one to remember the question |
| Interpretation of each major finding | Explain what each finding means in light of the research question | Findings are presented with no explanation of significance |
| Comparison to prior literature | Situate the study among existing research | The dissertation reads as if it exists in isolation |
| Confirm, extend, or contradict | State explicitly how findings relate to prior work, with reasoning | Comparison is vague, with no clear stance |
| Implications for practice, education, or policy | Say what should change as a result | Implications are generic and could apply to any study |
| Limitations | State what the design and sample cannot support | Limitations are boilerplate, disconnected from actual results |
| Recommendations for future research | Point toward the next study | No clear next step is offered |
The opening of the discussion chapter should remind the reader of the purpose and summarize the key findings in a paragraph, not a page. This is a reorientation, not a second results section. One or two sentences per major finding is usually enough before you move into interpretation.
The heart of the chapter is the section where each major finding is placed next to what the existing literature already said, and the relationship between the two is stated plainly. This is where a dissertation shows it understands its own contribution.
For each finding, decide which of three relationships applies, and say so directly rather than leaving the reader to infer it.
A discussion chapter that only ever confirms prior work, with no extension or contradiction anywhere, is a signal to many committees that the interpretation has not gone deep enough. Real studies usually produce at least one finding that does not simply repeat the literature.
Every dissertation eventually produces a finding that surprises the researcher: a relationship that was hypothesized but did not appear, or a result that runs opposite to what the literature predicted. How this is handled says a great deal about the rigor of the interpretation.
Do not ignore an unexpected finding by simply omitting it from the discussion, and do not force an explanation that the data cannot support. Instead, offer plausible, evidence-grounded reasons for the result. Consider sample characteristics that might explain it, such as a narrower age range or setting than prior studies used. Consider measurement limitations, such as an instrument that may not have captured the construct as precisely in this population. Consider that a null result can itself be informative, particularly if the study was adequately designed to detect an effect and did not find one, which may genuinely add to the literature rather than detract from it.
State the most plausible explanation, acknowledge that it is an interpretation rather than a proven cause, and move on. A paragraph that spends too long defending an unexpected result usually reads as more anxious than a shorter, more confident one that simply states the finding and offers a grounded explanation.
Share your results chapter, literature review, and program handbook with your brief, and a specialist can help you interpret findings, structure the chapter, or edit a full draft. The price is shown before you pay, and every delivered paper includes 14 days of free revisions.
A weak limitations section is one of the fastest ways to lose credibility in an otherwise solid chapter, because it signals that the writer either does not understand their own study's boundaries or is unwilling to state them. A strong limitations section does the opposite: it shows the researcher understands exactly what the study can and cannot claim.
A generic limitations section lists things that could be said about almost any study: small sample size, self-report bias, single-site design, potential for social desirability. These may all be true, but stated without connection to the actual study, they read as boilerplate rather than analysis. A strong limitations section names the specific characteristic of this study, explains why it matters, and connects it directly to how it affects interpretation of the actual findings.
| Generic limitations statement | Specific limitations statement |
|---|---|
| The sample size was small, which may limit generalizability. | The sample of 42 nurses from a single medical-surgical unit limits generalizability to nurses in other specialties or settings, and the modest sample size reduces power to detect smaller effects, which may explain the non-significant relationship between the two secondary variables. |
| Self-report measures may have introduced bias. | Burnout was measured using a self-report instrument completed during a shift, which may have been influenced by that day's workload rather than a stable trait, and could account for the wider variability seen in the afternoon-shift responses. |
| The study was conducted at one site, limiting generalizability. | Because the study was conducted at one academic medical center with an established shared-governance structure, the findings may not generalize to community hospitals or settings without similar structures already in place. |
Notice that each specific version does two things the generic version does not: it names the exact feature of the study, and it explains what that feature does to the interpretation of a particular result, not just to the study in the abstract.
Implications are the section most likely to feel like filler, and readers can usually tell when they are. An implication is earned when it traces directly back to a specific finding from this study. It is not earned when it is a general statement about the importance of the topic that could have been written before the study was ever conducted.
| Not earned (generic) | Earned (traces to a finding) |
|---|---|
| This study highlights the importance of addressing nurse burnout in hospital settings. | Because burnout scores were highest among nurses working more than three consecutive twelve-hour shifts, unit scheduling practices that limit consecutive shift length may be a more targeted lever than general wellness programming. |
| Nursing education should continue to emphasize communication skills. | Because new graduates in this sample underestimated their own handoff completeness relative to observed practice, orientation programs may benefit from structured, observed handoff feedback rather than self-assessment alone. |
Write each implication as a sentence that begins with the finding and ends with what should change, and the connection stays visible to the reader instead of being asserted from nowhere.
The recommendations for future research section closes the interpretive work of the chapter by pointing to what the study could not answer. Committees read this section closely, because it shows whether the researcher understands the boundaries of their own work well enough to say what should come next, and it is one of the more common places a defense question originates.
The strongest recommendations trace directly back to either a limitation stated earlier in the chapter or a gap the findings exposed, rather than restating the broader research agenda of the field. A recommendation that could have been written before the study was conducted, with no reference to what this particular study found or could not answer, is too generic to be useful.
| Generic recommendation | Recommendation tied to this study's limits |
|---|---|
| Future research should explore nurse burnout further. | Because this study's cross-sectional design could not establish whether scheduling patterns precede or follow burnout symptoms, a longitudinal design following nurses across a full rotation cycle would clarify the direction of the relationship. |
| More studies are needed on patient education interventions. | Because this study measured only immediate post-teaching comprehension, a follow-up study measuring retention and actual self-management behavior at thirty and ninety days would test whether the gains observed here persist. |
Keep these two lists separate in the chapter. Recommendations for future research describe what a subsequent study should investigate, using terms like design, sample, measurement, or setting. Recommendations for practice, covered under implications above, describe what clinicians, educators, or administrators should do differently now, based on what is already known from this study. Mixing the two together in one list tends to blur both, since a reader has to work out which kind of recommendation each item actually is.
Illustrative example, not a real client. This short story is invented to show the pattern, and it contains no real people or data.
The problem. A PhD candidate had a completed results chapter with a non-significant relationship between two of her four hypothesized variables, and her draft discussion chapter simply stated the result and moved on.
The tension. Her committee returned the draft asking why the discussion chapter did not engage with the non-significant finding at all, and noted that her limitations section read identically to a template she had clearly copied.
The turn. She went back to her sample characteristics and noticed the non-significant pair had the narrowest range of scores in her entire dataset, which gave her a plausible, defensible explanation grounded in her own data rather than a guess.
The proof. She rewrote the limitations section to name that restricted range specifically, rather than listing generic study limitations, and rewrote the interpretation of the non-significant finding around it.
The payoff. Her committee's next round of feedback focused on formatting and a few citation details rather than the substance of her interpretation.
The discussion chapter is where the dissertation's argument comes together. The introduction posed a question, the literature review showed what was already known, the methods chapter explained how the question would be answered, and the results chapter reported what was found. The discussion chapter is the only place where all of that is drawn together into a coherent statement of what the study contributes, what it does not settle, and what should happen next.
A discussion chapter that does this well leaves the reader with a clear sense of the study's place in the field: what it adds, what it does not claim, and where the next researcher should look. That clarity is usually a better predictor of a smooth defense than any individual finding in the results chapter. See our dissertation defense preparation guide for how this chapter typically comes up in defense questions.
Length depends on your program and the number of findings you have to interpret. Follow your handbook, and prioritize thorough interpretation of each major finding over reaching a target page count.
Often yes, particularly literature that helps explain an unexpected finding. Check your program's expectations, since some committees prefer all cited literature to appear in chapter two, with the discussion chapter only referring back to it.
A study with entirely non-significant findings can still make a genuine contribution, particularly if it was adequately designed and the null result is informative. Interpret why the results turned out that way using your sample, measures, and setting, rather than treating the chapter as a failure to explain away.
The same principle applies: the findings chapter reports themes, and the discussion chapter interprets what those themes mean, how they relate to prior qualitative or quantitative literature on the topic, and what they imply. The comparison to literature and the limitations section still apply, framed around trustworthiness rather than statistical generalizability.
Most programs expect a full draft, including the discussion chapter, before the defense. Some chairs prefer to review the results chapter first and give feedback before you draft the discussion. Ask your chair directly rather than assuming either sequence.
A strong discussion chapter does three things well: it explains what the findings mean, it says exactly what the study cannot claim, and it points somewhere useful for the field to go next. Keep every interpretation tied to an actual finding, keep the limitations specific to your own study, and the chapter will do the work a dissertation's final argument needs it to do.
Want help interpreting your findings or drafting this chapter? Get my instant quote. The price is shown before you pay, every delivered paper includes 14 days of free revisions, and refund terms are on the money-back guarantee page. Please use any model paper in line with your institution's academic-integrity policy.