How to Write the Discussion Chapter: Structure, Interpretation, Limitations and Recommendations

A recommended structure for the discussion of a seminar paper or thesis - opening, discussion by hypothesis, unexpected findings, limitations, implications and future research - with example wording and common mistakes.

Part of the series: Seminar Paper Structure: Every Chapter, What Goes In It and How Long It Should Be

The discussion is where the paper turns from research into insight. In the results you said what was found; in the discussion you answer what it means. It's also the easiest chapter to write badly: repeating the results in other words, or copying the introduction. The structure below prevents both.

The recommended structure

1. Opening: what you tested and what emerged (one paragraph)

One sentence on the aim, and two or three on the main findings - in words, without statistics. "The study examined the relationship between social support and burnout among social workers. Higher support was related to lower burnout, a relationship partly explained by self-efficacy."

2. Discussion by hypothesis

A paragraph or two per hypothesis (or main finding): - What was found - briefly. - How it fits the literature from your introduction: consistent (who found similar?) or contradictory (who found otherwise, and why might that be?). - The explanation - a possible mechanism, a supporting theory.

"This is consistent with earlier studies showing... (Cohen, 2019). However, unlike Miller's (2020) findings among teachers, the relationship here was strong. This may be because..."

3. Unsupported hypotheses and unexpected findings

Don't hide them and don't apologise. Offer possible explanations: sample size, measurement, population characteristics, or the possibility that the theory itself needs updating. A well-explained unsupported hypothesis is a legitimate, even interesting, finding.

4. Limitations

Three or four real limitations, each with its consequence: - Sample: "A convenience sample from one district calls for caution in generalising." - Design: "A cross-sectional correlational design does not allow causal conclusions." - Measurement: "Self-report questionnaires are open to social-desirability bias."

Don't list limitations unrelated to the findings, and don't lead with them.

5. Implications and future research

• Practical: what an organisation, therapist or teacher can do with the finding.
• Theoretical: what it adds to understanding in the field.
• Future research: suggestions that follow from the limitations (longitudinal design, another population, objective measures).

6. Conclusion (one paragraph)

The paper's central message in three or four sentences - what the reader should remember.

Length

Seminar paper: 4-6 pages, about as long as the results or a little longer. Thesis: 10-20 pages.

Common mistakes

• Repeating the results with the numbers - they're already in the results chapter.
• Bringing in brand-new sources as the main basis of the discussion. A new source to explain an unexpected finding is fine, in moderation.
• Over-generalising: "the study proves that..." - one study doesn't prove; write "suggests", "supports".
• Causal interpretation of a correlational study.
• Generic limitations ("the sample could have been larger") without saying how they affect the conclusions.
• A discussion that doesn't return to the hypotheses in a clear order.

FAQ

Discussion and conclusions - one chapter or two? It depends on the guidelines; in a seminar paper usually one, with a concluding paragraph.

First person? APA allows it, but check your supervisor's preference.

Where does it fit? It's the chapter after the results: Seminar paper structure; for writing the results themselves: Statistical analysis for a seminar paper. Want feedback on your discussion draft? Send it to us.

Last updated: 2026-10-01