How to Present Results in a Research Paper (Clear, Concise, Convincing)

A well-written results section does something deceptively simple: it tells the reader exactly what you found. No interpretation. No speculation. No spin. Just the findings, presented clearly enough that a reader could draw their own conclusions before you discuss yours.

That simplicity is exactly what makes it hard. We've edited results sections where three paragraphs of text said less than a single well-designed table. We've seen researchers bury their most important finding in the middle of a paragraph about preliminary analyses. We've watched numbers contradict the text that describes them — sometimes in the same sentence.

Learning how to write a results section for a research paper means learning restraint. Here's what that looks like in practice.

The cardinal rule: results vs. discussion

Before anything else, understand the boundary. Your results section reports what happened. Your discussion section explains what it means. Mixing the two is the fastest way to weaken both sections.

Results territory: "Participants in the experimental group scored 23% higher than the control group (M = 78.4, SD = 12.1 vs. M = 63.7, SD = 14.3), t(198) = 7.42, p < .001, d = 1.05."

Discussion territory: "This effect size exceeds those reported in similar interventions (Smith, 2022; Lee, 2023), suggesting our modified protocol may be more effective than the standard approach."

See the line? The results section gives you the numbers. The discussion section tells you what the numbers mean in context. When you catch yourself writing "this suggests" or "this may indicate" in your results — stop. Move that sentence to the discussion.

There are exceptions. Some journals and some disciplines blend results and discussion into a single section. If your target journal does this, follow their format. But even in combined sections, keep the reporting and interpreting clearly distinguished within each paragraph.

How to structure your results section

Organization matters more in results than in any other section. Your reader is processing numbers, and numbers without structure become noise.

Option 1: Follow your research questions. If your paper poses three research questions, organize your results into three sections that answer them in order. This creates a direct mapping between what you asked and what you found — easy for the reader, easy for the reviewer.

Option 2: Follow your hypotheses. Similar to above, but structured around predictions rather than questions. For each hypothesis, present the relevant data and state whether the hypothesis was supported. Be direct: "Hypothesis 1 was supported" or "Hypothesis 2 was not supported."

Option 3: Follow analytical logic. Start with descriptive statistics and preliminary analyses (reliability, normality checks, correlation matrices). Then move to primary analyses. Then secondary or exploratory analyses. This structure works well for complex studies with multiple analytical steps.

Whichever option you choose, open your results section with a brief orienting paragraph. Tell the reader how this section is organized: "We first report descriptive statistics and preliminary analyses, followed by the results of our primary regression models, and finally our exploratory mediation analysis." This roadmap saves your reader from wondering where you're headed.

Presenting research results: text, tables, and figures

You have three tools for presenting data. Using the wrong one is like using a hammer on a screw — it technically works, but the result is ugly.

Text is best for simple findings with few numbers. "The mean age of participants was 34.2 years (SD = 8.7), and 62% identified as female." If a finding involves one or two numbers, put it in the text.

Tables are best for precise comparisons across multiple groups or conditions. If you have more than three numbers to compare, a table is almost always clearer than text. Tables are also the right choice when exact values matter — when someone might want to reference your specific means, standard deviations, or p-values.

Figures are best for showing patterns, trends, and distributions. If the relationship matters more than the exact numbers — a growth curve, a comparison of distributions, an interaction effect — use a figure. Figures are processed faster than tables and they stick in memory longer.

The critical rule: never repeat data across formats. If a finding appears in a table, don't describe every number from the table in the text. Instead, highlight the key pattern: "As shown in Table 2, treatment effects were strongest in the high-dosage condition." The text directs attention. The table provides detail.

We see researchers violate this rule constantly. The results section describes every cell of every table in paragraph form. This doubles the length without adding information. Your text should interpret the table's story, not recite its contents.

Writing about statistical results clearly

Statistical reporting has conventions, and following them signals competence to reviewers.

Report effect sizes, not just significance. A p-value tells you whether an effect exists. An effect size tells you whether it matters. Report both. "The intervention group outperformed controls, t(198) = 7.42, p < .001, d = 1.05" — the d = 1.05 tells the reader this is a large effect, which matters far more than the p-value for practical purposes.

Be consistent in notation. Pick a reporting format and stick with it throughout. If you report means as "M = 78.4" in paragraph one, don't switch to "the mean was 78.4" in paragraph three. Consistency signals attention to detail.

Round appropriately. Two decimal places for most statistics. Three for p-values when they're very small (p = .002). Never report p = .000 — write p < .001 instead. These small details matter to reviewers who read results sections all day.

Lead with the finding, not the analysis. "Participants who received the intervention scored higher on the creativity measure" tells the reader the result. "A one-way ANOVA was conducted on creativity scores" tells the reader the method. Lead with the finding: "Intervention participants scored significantly higher on creativity (M = 42.3, SD = 8.1) than controls (M = 35.7, SD = 9.2), F(1, 196) = 28.41, p < .001, η² = .13."

Common results section mistakes

Interpreting results in the results section. We said it above but it bears repeating — this is the single most common mistake. Save your interpretation for the discussion section. Your results section is a courtroom transcript, not closing arguments.

Burying key findings. Your primary finding should appear in the first substantive paragraph after any preliminary analyses. Don't make the reader hunt for it behind demographic descriptions and reliability statistics. Report the main finding first, then the secondary findings, then the exploratory ones.

Ignoring non-significant results. If you tested a hypothesis and it wasn't supported, report that. Selectively reporting only significant results is a form of bias that distorts the scientific record. "There was no significant difference between groups on measure Y, t(198) = 0.87, p = .384, d = 0.12" — report it cleanly and move on.

Over-describing tables and figures. If Table 3 shows means and standard deviations for all six conditions, you don't need a paragraph describing each value. Write: "Table 3 presents descriptive statistics for all conditions. The highest scores appeared in Condition A, while Condition F showed the greatest variability." Direct the reader's attention. Let the table do the heavy lifting.

Inconsistent formatting. Mixed reporting styles, inconsistent decimal places, and switching between APA and non-APA notation within the same section. Run your results through our AI proofreader to catch these inconsistencies — they're nearly invisible to the writer but glaringly obvious to reviewers.

The AI summarizer can also help during the revision process. If your results section runs long, feed it into the summarizer to identify which passages contain genuine findings and which are redundant descriptions. Trim accordingly.

AI Proofreader for Academic Papers

Catch inconsistent statistical notation, grammar errors, and formatting issues. Designed for the precision academic writing demands.

Frequently asked questions

Q: How long should a results section be?

Your results section should be as long as it needs to be to report all findings — and no longer. For a typical journal paper with two or three research questions, 800–1,500 words plus tables and figures is common. The key metric isn't word count but information density: every paragraph should report at least one finding. If a paragraph contains only methodological description or transition language, it's padding.

Q: Should I report all results or only significant ones?

Report all results — significant and non-significant — for every hypothesis or research question you tested. Selective reporting of only significant results is considered a questionable research practice and violates APA reporting standards. Non-significant findings are informative. They tell the field where effects don't exist, which prevents other researchers from testing the same dead ends.

Q: When should I use a figure vs. a table in my results?

Use tables when exact values matter and when readers might want to reference specific numbers. Use figures when patterns, trends, or relationships matter more than precise values. A good rule: if someone would need to squint at your figure to extract a specific number, that data belongs in a table. If someone would struggle to see a pattern by scanning columns of numbers, that data belongs in a figure.

Q: Can I include raw data in my results section?

No — your results section presents analyzed data, not raw data. Include descriptive statistics (means, standard deviations, frequencies) and inferential statistics (test statistics, p-values, effect sizes). Raw data belongs in supplementary materials or a data repository, with a note in your methods section about where it can be accessed. Some journals now require data availability statements, which you should check before submission.