How to Write a Discussion Section (With Examples & Structure)

The discussion section is where most students either shine or fall apart. It's the part of your paper where you stop reporting numbers and start making sense of them. What do your findings actually mean? Do they agree with what other researchers found? What are the real-world implications?

If you've ever gotten feedback like “you're just repeating your results” or “where's the interpretation?”, this guide is for you.

I'll walk you through what the discussion section actually does, the five parts every good discussion includes, how to structure it, and the mistakes that trip students up the most. No jargon. No hand-waving. Just the stuff that works.

(If you haven't written your methodology section yet, start there. The discussion comes after your results.)

What Does a Discussion Section Do?

The discussion section has one job: interpret your results and explain what they mean in the context of your field.

That sounds simple, but it trips people up because it requires a different kind of thinking than the rest of the paper. In the Results section, you present data objectively. In the Discussion, you step back and think critically. You connect the dots between what you found and what everyone else has found. You explain the “so what.”

Think of it this way: if your Results section is a photograph, your Discussion is the caption that explains what the photo shows and why it matters.

A strong discussion does four things:

• Interprets your key findings (not just restates them)
• Connects your results to existing research
• Explains the practical and theoretical implications
• Acknowledges limitations and suggests where research should go next

If your discussion does all four, you're in excellent shape.

What Is the Difference Between Results and Discussion?

This is the confusion that causes most of the problems. A lot of students blur the line between Results and Discussion, either by interpreting too much in their Results section or by repeating raw data in their Discussion.

Here's the cleanest way to think about it:

See the difference? Results report numbers. Discussion explains what those numbers mean, why they matter, and how they connect to other research. The Results section says “here's what happened.” The Discussion says “here's why that's interesting.”

What Are the 5 Parts of a Discussion Section?

Every good discussion section covers five things. You don't always need five separate subheadings, but the content should be there. Think of these as the building blocks.

1. Interpretation of Key Findings

Start by restating your most important findings, briefly , and then immediately explain what they mean. This is not the place to dump your entire Results section again. Pick the two or three findings that matter most and dig into them.

Ask yourself: why did I get these results? What mechanisms might explain them? Were they expected or surprising?

A good structure for each finding is: state the finding in one sentence, then spend two to four sentences interpreting it. Explain the possible reasons behind it. If you had a hypothesis, say whether the finding supports or contradicts it and why.

2. Comparison With Existing Literature

This is what separates a mediocre discussion from a strong one. You need to connect your findings to what other researchers have published.

For each major finding, ask: does this agree with prior research? If yes, say so and cite the relevant studies. If your results contradict existing findings, that's even more interesting, explain why there might be a discrepancy. Different sample? Different methodology? Different context?

This is where your literature review pays off. If you did a thorough lit review earlier in your paper, you already know the key studies to reference. Pull them back in here and show how your work fits into (or challenges) the bigger picture.

3. Implications (Theoretical and Practical)

Implications answer the question: so what? Why should anyone care about your findings?

There are two types. Theoretical implications are about what your findings mean for the existing body of knowledge. Do they support an existing theory? Challenge it? Extend it to a new context?

Practical implications are about real-world applications. If you studied a teaching method, what should teachers do differently? If you studied a drug interaction, what should clinicians know? If you studied consumer behavior, what should marketers take away?

Don't overpromise here. Be specific and grounded. Instead of “this could change education forever,” write “these findings suggest that incorporating spaced repetition into weekly lesson plans could improve test performance in undergraduate courses.”

4. Limitations (and How to Present Them Honestly)

Every study has limitations. Your supervisor knows this. Your reviewers know this. Pretending your study is perfect doesn't make you look confident, it makes you look naive.

The key is to present limitations honestly but without undermining your entire paper. State the limitation, explain how it might have affected your results, and (where possible) explain what you did to mitigate it or what future research could do differently.

Common limitations to address:

• Sample size or composition: “The sample was limited to undergraduate students at a single university, which may limit generalizability.”
• Methodological constraints: “Self-reported data may be subject to social desirability bias.”
• Scope: “This study focused on short-term effects; long-term outcomes were not assessed.”
• External factors: “Data collection occurred during COVID-19, which may have influenced participant behavior.”

5. Future Research Directions

This is your chance to show that you understand the bigger picture and that your work opens doors for others. Based on your findings and limitations, what should future studies investigate?

Be specific. “More research is needed” is the most overused sentence in academia and it tells the reader nothing. Instead, say exactly what kind of research, with what kind of sample, using what methods, and why.

Good future research suggestions typically flow directly from your limitations. If your sample was small, suggest a larger-scale replication. If you only looked at short-term effects, suggest a longitudinal study. If you studied one demographic, suggest cross-cultural research.

How Should You Structure a Discussion Section?

The most effective discussion sections follow what I call the “inverted funnel” structure. You start narrow , with your specific findings, and gradually zoom out to broader implications.

Here's the flow:

Notice how you move from your specific data points outward to the broader research landscape, then to real-world impact, and finally to what still needs to be done. That progression gives the reader a sense of closure and completeness.

Some disciplines combine Results and Discussion into one section. If that's the case for your paper, the same structure applies, you just weave interpretation in right after presenting each result, rather than saving it all for a separate section.

Step-by-Step Writing Process

Here's how to actually sit down and write the thing. This process works whether your discussion is 800 words or 3,000.

Step 1: List Your Key Findings

Before you write a single paragraph, open your Results section and pull out the two to five most important findings. Write each one as a single sentence. These are your discussion anchors, everything else revolves around them.

Step 2: For Each Finding, Ask “Why?”

Under each finding, jot down possible explanations. Why did this happen? What theory supports this? Was it expected? If it was unexpected, what could account for the surprise? Don't worry about writing polished prose yet, just get your reasoning down.

Step 3: Connect Each Finding to the Literature

Go back to your literature review and identify which studies relate to each finding. Note whether your result agrees or disagrees with each one, and write a sentence or two about why. This step is much faster if you kept good notes during your lit review.

Step 4: Draft the Interpretation Paragraphs

Now write. For each major finding, aim for one to three paragraphs that cover the interpretation and the literature comparison. Use this formula: state the finding briefly, interpret it, compare it with existing research, and explain any discrepancies.

Step 5: Write the Implications

Ask yourself: what should someone do differently because of my findings? Write one paragraph on theoretical implications and one on practical implications. Be concrete. Vague implications are as bad as no implications.

Step 6: Write the Limitations and Future Directions

List three to five limitations. For each one, write one to two sentences explaining the limitation and how it might affect interpretation. Then turn each limitation into a future research suggestion. This section usually writes itself once you've been honest about what your study can and cannot claim.

Step 7: Revise for Flow and Redundancy

Read through the whole discussion from start to finish. Check that each paragraph connects logically to the next. Cut any sentences that just repeat your Results without adding interpretation. Make sure you're not introducing new data, the Discussion is for interpretation, not new findings.

What Mistakes Should You Avoid in a Discussion Section?

These are the errors I see most often. Avoid them and your discussion will already be stronger than most.

1. Repeating your results verbatim

This is the number one problem. If your Discussion reads like a copy of your Results section with “this is important” tacked on the end, it needs a rewrite. You can reference findings, but the emphasis must be on interpretation, not repetition.

2. Introducing new data

The Discussion is not the place to present results you forgot to include earlier. If it's data or a statistical test, it belongs in the Results section. The Discussion is exclusively for interpretation and meaning.

3. Being too defensive about limitations

Some students write their limitations section like an apology letter. “This study is limited because… and also limited because… and therefore all conclusions should be taken with extreme caution.” That undermines your entire paper. State your limitations matter-of-factly, explain how they might affect the results, and move on.

4. Ignoring contradictory findings

If your results contradict a well-known study, don't pretend that study doesn't exist. Engage with the contradiction. Possible explanations, different methods, different populations, different contexts , actually make your discussion more interesting and credible. Ignoring contradictions makes it look like you didn't do your homework.

5. Making claims your data doesn't support

Be careful with your language. If you ran a correlational study, you cannot claim causation. If your sample was 30 undergrads, you cannot generalize to “all humans.” Use hedging language where appropriate: “suggests,” “indicates,” “may,” “it is possible that.” Your supervisor will catch overstatements every time.

How Long Should a Discussion Section Be? (Word Count Guide)

There is no universal rule, but here are realistic benchmarks based on what supervisors and journals typically expect:

As a rough guide, your Discussion should be about 25 to 35 percent of your total paper length. It's typically one of the longest sections alongside the Literature Review.

If you're writing a thesis and your discussion is under 1,500 words, you're probably not interpreting deeply enough. If it's over 3,000 words, check whether you're repeating yourself or including content that belongs in a different section.