Writing a research report: chapter structure, quantitative results and clear prose
Readers should be able to identify your question, methods, findings and contribution without struggling through inflated prose. Short, clear sentences and enough statistical detail to assess the analysis serve a dissertation better than unnecessary length.
A research report communicates an investigation. Each section has a task: introduce the question, establish its scholarly basis, explain the methods, report findings, interpret them and draw proportionate conclusions. Quantitative reporting should allow scrutiny and replication. Prefer clear sentences, consistent tense and an identifiable actor over language chosen merely to sound academic.
1. Write throughout the project
Writing clarifies thinking; it is not just the final stage. Proposals, notes, reflexive journals, literature-review drafts and methods chapters develop alongside collection and analysis (Saunders et al., 2023, pp. 718–720). Familiarity with a topic can make it harder to recognise what a new reader needs explained.
- Write regularly, even for a focused 15–30 minutes.
- Reserve your most alert time for demanding writing.
- Outline chapters and give each paragraph a topic sentence.
- Finish at a useful stopping point and reread before continuing.
- Back up current and earlier versions daily.
- Seek candid feedback before submitting to your supervisor.
Suggestions adapted from Saunders et al. (2023, pp. 721–723).
2. Choose a reporting structure
Structure foregrounds particular aspects of a study. Saunders et al. (2023, pp. 724–725), drawing on Yin, discuss five approaches.
| Approach | Emphasis | Suitable for |
|---|---|---|
| Linear-analytic | Logical sequence of the investigation | Deductive, theory-testing work; the conventional structure |
| Comparative | Compare datasets or alternative perspectives on one dataset | Multiple cases or groups |
| Chronological | Events and their context over time | Historical or change-process research |
| Theory-building | Progressive development of themes and explanations | Inductive or grounded-theory work |
| Suspense | Present the answer first, then examine competing explanations | Explanatory studies with rival hypotheses |
Check that the institution permits the format, readers can follow it and it shows how the question was answered. Quantitative dissertations commonly use a conventional linear-analytic structure (Saunders et al., 2023, p. 725).
3. What each conventional chapter should do
| Section | Task | Source pages |
|---|---|---|
| Abstract | Question and importance, methods, findings and conclusion | 726 |
| Introduction | Problem, rationale, question, aim, objectives and relevant context | 726 |
| Literature review | Theoretical context leading to the question and hypotheses | 727–728 |
| Methods | Enough transparency to assess the findings | 729–730 |
| Results | What the analysis found | 730–732 |
| Discussion | Meaning, relation to theory, strengths and limitations | 732–734 |
| Conclusion | Answer, contribution, implications and future research; no new analysis | 734–735 |
| References and appendices | Complete source matching and supporting materials | 735–736 |
A common abstract length discussed by Saunders et al. (2023, p. 726).
A usual APA abstract limit, subject to the intended publication's instructions (APA, 2020, Section 2.9).
Words in a multi-idea sentence the textbook uses to illustrate the need for revision (Saunders et al., 2023, p. 747).
Abstract and introduction
An abstract should stand alone, reflect the report accurately and be concise, objective and readable. Drafting one early can clarify the argument; revise it after completing the report. Do not exaggerate or include findings absent from the report (Saunders et al., 2023, pp. 726–727).
Literature review
Use an informative chapter title where permitted. Distinct bodies of literature may require more than one chapter. The critical-review article explains thematic structure and synthesis in detail (Saunders et al., 2023, p. 728).
Methods
Explain exactly how evidence was collected and why, using methodological sources to justify choices. Readers need enough detail to assess conclusions and repeat the work. This includes limitations: omitting a weakness does not make an analysis reproducible (Saunders et al., 2023, pp. 729–730; Cooper, 2020, pp. 4, 31).
Results and discussion
Separate “what I found” from “what I infer”. The first belongs in results; the second in discussion and conclusions. Results may follow objectives or themes if the organisation is clear (Saunders et al., 2023, pp. 730–731).
Discussion answers “so what?”. State which hypotheses were supported, compare findings with earlier work and explain theoretical and practical meaning. Address bias, measurement precision, multiplicity, sample adequacy and generalisability. Specific limitations demonstrate understanding rather than failure (APA, 2020, Section 3.8, pp. 89–90; Saunders et al., 2023, pp. 732–734).
Conclusion, recommendations and appendices
Answer the question without introducing new material and connect the contribution with the review. If recommendations are an intended output, include them in the objectives. Appendices support rather than inflate the report; keep material essential for understanding in the main text, while including relevant instruments and information or consent materials as appropriate (Saunders et al., 2023, pp. 734–736).
4. Reporting quantitative findings
Report relevant findings, including nonsignificant and contradictory results. For regression or SEM, describe the model estimated, software and version, estimation problems such as nonconvergence, and assumption violations affecting interpretation (APA, 2020, Section 3.7, pp. 86–89).
APA's numerical guidance includes appropriate precision, commonly two decimal places for correlations and test statistics, one decimal place for means and standard deviations of integer-scale measurements where adequate, and exact p values to two or three decimals. Values below .001 are reported as p < .001 rather than p = .000. Apply precision to the context rather than copying raw software output (APA, 2020, Section 6.36, p. 180).
Report effect sizes and confidence intervals because a p value does not quantify magnitude. Give enough information to identify the analysis, including degrees of freedom where relevant. If descriptive values appear in a table, highlight the meaningful pattern instead of repeating every cell (APA, 2020, Sections 3.7 and 6.43).
APA's Journal Article Reporting Standards include guidance specific to structural equation modelling and Bayesian methods. Consult the appropriate standards rather than relying only on the general checklist (Cooper, 2020, p. 5).
English APA examples use decimal points and omit a leading zero for statistics that cannot exceed one, such as p and r. A Vietnamese institution may require decimal commas. Follow its requirements consistently across prose and software tables.
Tables, figures and commentary
Keep central results in the main report. Refer to tables before they appear and explain the important pattern. Use tables for repeated information; a few values may be clearer in prose (Saunders et al., 2023, pp. 732, 746).
5. Build a clear storyline
A useful title communicates the content with few words. Ask another person what they expect from the title alone. An early abstract also tests whether the argument can be stated simply (Saunders et al., 2023, p. 744).
Try reasoning backwards: state the conclusion, explain the findings supporting it, then the methods producing those findings, and finally the design, question and literature justifying the methods. A break in that chain reveals an argument that needs work (Saunders et al., 2023, pp. 744–745).
Drawing on Lange and Pfarrer, the book identifies common ground, complication, concern, course of action and contribution as building blocks of an academic storyline. Consistent headings and brief chapter openings and endings help readers maintain their place (Saunders et al., 2023, pp. 745–746).
6. Clear prose before impressive prose
Clear writing exposes weaknesses in reasoning; obscure prose can hide a useful idea. Prefer an understandable explanation over language selected to sound academic (Saunders et al., 2023, pp. 746–747).
One main idea per sentence
Long sentences often bury their main point in subordinate clauses. Read aloud, identify the separate ideas and split where needed. The book rewrites a 101-word sentence containing at least six ideas into shorter sentences using approximately half the words (Saunders et al., 2023, p. 747).
Terminology, jargon and quotations
Necessary technical terminology belongs in a research report; empty jargon does not. Define unfamiliar terms and use a glossary where helpful. Avoid constructing a chapter from connected quotations: paraphrase accurately and show your understanding while retaining attribution (Saunders et al., 2023, p. 748).
Person, tense and voice
Impersonal past-tense passive writing is not a universal requirement. First person can be appropriate when the researcher is part of the process. APA permits both active and passive constructions, recommending an identifiable actor where it matters (Saunders et al., 2023, p. 749; APA, 2020, Section 4.13).
| Section | Usual tense | Example |
|---|---|---|
| Earlier studies in the review | Past; present perfect for continuing research | “Nguyen (2021) reported…”; “Studies have used…” |
| Methods | Past | “Participants completed the questionnaire.” |
| Results | Past | “H1 was supported.” |
| Discussion of meaning | Present | “The findings indicate…” |
| Conclusions, limitations and directions | Present | “This study has several limitations.” |
Vietnamese does not inflect verbs as English does, but temporal markers and sentence construction should remain consistent. Abrupt tense shifts can confuse readers (APA, 2020, Section 4.12).
Bias-free language, confidentiality and revision
Avoid assuming a group's gender. Protect promised confidentiality in names, descriptions and appendices. Revise in several rounds, allow a break between drafts and repeatedly ask what would make the argument easier to follow (Saunders et al., 2023, pp. 749–750).
7. Assessment and the defence
Using Bloom's taxonomy, Saunders et al. (2023, p. 751) distinguish understanding, application and analysis from synthesis and evaluation. Methods demonstrate application; analysis examines relationships; the storyline and conclusions synthesise; the review and a realistic limitations section evaluate. The report must also meet format, length and required-section rules.
Ask whether you would be comfortable seeing the current version in the university library. If not, identify the most important unfinished work without polishing one section indefinitely and missing the deadline (Saunders et al., 2023, pp. 751–752).
For the defence, establish the format, presentation and question time, and audience in advance. Prepare a coherent central message that remains memorable after the presentation (Saunders et al., 2023, p. 753).
8. Templates, common mistakes and checklist
Separating findings from interpretation
| Question | Results: what was found | Discussion/conclusion: what it means |
|---|---|---|
| [Question 1] | Estimates, uncertainty and test results | Meaning, comparison with earlier work and implications |
| [Question 2] | [Evidence] | [Interpretation] |
Illustrative regression-reporting paragraph
Table [number] reports estimates for [model] using [n] observations and [software/version]. Assumptions concerning [residuals, collinearity and heteroscedasticity] were assessed and [state findings and remedies]. [A] was positively associated with [B], β = [value], t([df]) = [value], p = [exact value], 95% CI [[lower], [upper]], supporting H[number]. The estimate for [C] was not statistically significant, β = [value], p = [value], and H[number] was not supported. The model's R² was [proportion], corresponding to [100 × R²]% of observed variation in [B].
| Weak approach | Problem | Better approach |
|---|---|---|
| Comparing earlier studies within a separate results chapter | Findings and interpretation become confused | Use the discussion for meaning and comparison |
| Reporting only supported hypotheses | Selective omission | Report all relevant findings |
| Only “p < .05”, with no effect size | Magnitude is unknown | Give suitable exact p values, estimates and intervals |
| Main results only in appendices | Central evidence becomes hard to assess | Place key tables in the main text |
| Long sentences containing many ideas | The argument is buried | Separate ideas and read aloud |
| New analysis in the conclusion | No earlier evidential foundation | Move the analysis to results and synthesise at the end |
| Generic apologies for limitations | Readers cannot assess scope | Name specific biases, measures, sample limits and inference boundaries |
Draft checklist
Condensed from Saunders et al. (2023, p. 751).
- Structure, title and storyline are clear.
- The abstract reflects the report and the introduction states the question and objectives.
- The review leads to the study and the methods are justified.
- Findings and interpretations have identifiable places.
- Chapters have suitable openings, endings and headings.
- Prose is clear, unnecessary jargon removed, and language checked.
- References are complete and supporting material is appropriately placed.
- Institutional formatting and submission requirements are met.
Conclusion
A sound analysis still needs a readable report. Distinguish evidence from commentary, provide enough detail for scrutiny and make the argument easy to follow. The decisions made throughout the project reappear in the final document; good notes and transparent records make that final writing much easier.
References
Main source
- Saunders, M. N. K., Lewis, P., & Thornhill, A. (2023). Research methods for business students (9th ed.). Pearson. Pearson
Reporting standards
- American Psychological Association. (2020). Publication manual of the American Psychological Association (7th ed.). https://doi.org/10.1037/0000165-000
- Cooper, H. (2020). Reporting quantitative research in psychology: How to meet APA Style Journal Article Reporting Standards (2nd ed.). American Psychological Association. https://doi.org/10.1037/0000178-000
Books are paraphrased with page and section references; secondary citations are identified. The sample regression paragraph and explanations of Vietnamese decimal and tense conventions were written for this article. Numerical placeholders are illustrative, not empirical findings.
