You finished running your analyses. SPSS or R gave you pages of output, and now your thesis document has a blank heading that reads "Chapter 4: Results." The cursor blinks. You know the numbers but you are not sure what goes here versus what belongs in the Discussion chapter. This confusion costs students weeks of rewrites. The results section has one job: report what you found, clearly and completely, without explaining why. This guide covers the structure, APA formatting rules, and the mistakes that send drafts back for revision.
What the Results Section Does (and Does Not Do)
The results section answers your research questions with data. It tells the reader what happened when you ran each test. Did the groups differ? Was there a correlation? Did the model predict the outcome? You present the statistical evidence, and you stop there.
Interpretation belongs in the Discussion chapter. If you catch yourself writing "this may be because..." or "a possible explanation is..." you have crossed the line. The results section does not explain causes, compare your findings to other studies, or cite previous literature. It also does not contain raw data dumps, SPSS screenshots, or long tables of individual responses. Those go in appendices, if anywhere.
Think of it this way: a reader should be able to read your results section and know every statistical outcome without forming any opinion about what those outcomes mean. A second reader should be able to read your Discussion and understand why those outcomes matter. The two chapters work as a pair, but they do not overlap. In roughly 85% of the thesis projects we review, the first draft mixes results with discussion. Separating them cleanly is the single fastest way to improve a results chapter.
The Three-Part Structure
Most results sections follow a predictable order. Not because there is a rigid rule, but because this sequence makes logical sense and matches what committees expect to see.
1. Preliminary Analyses
Open with descriptive statistics and assumption checks. Report the mean, standard deviation, and sample size for each variable or group. If you used a questionnaire, report its reliability here (Cronbach's α). A scale with α = .83 across 24 items tells the reader the measurement was internally consistent before you built any comparisons on it. If you ran normality tests, report them here too. A Shapiro-Wilk result of W(64) = .97, p = .142 confirms that the distribution did not deviate significantly from normal, which justifies using parametric tests in the next section. For details on how to run and interpret these checks, see our normality testing guide.
This section is usually short. Half a page to a full page. Its purpose is to establish that your data is clean and your assumptions are met before you test any hypotheses.
2. Main Analyses
Present your hypothesis tests in the same order you listed them in the methodology chapter. Hypothesis 1 first, then Hypothesis 2, and so on. Each hypothesis gets its own paragraph or subsection. If you had four hypotheses, the reader can scan the results section and find the answer to each one without hunting.
For each hypothesis, follow a four-sentence pattern: state the hypothesis, name the test you used, report the result, and state whether the hypothesis was supported or not supported. You can stretch this to five or six sentences when the result requires more context, but the underlying logic stays the same.
3. Additional or Exploratory Analyses
If you ran analyses beyond your stated hypotheses, report them last. Maybe you noticed an unexpected pattern and ran a post hoc comparison. Maybe your supervisor asked you to check for a moderating effect that was not part of the original research questions. Label this section clearly so the reader knows these findings were not pre-planned. Committees are fine with exploratory results as long as you do not dress them up as confirmatory.
What to Report for Every Statistical Test
Regardless of whether you ran a t-test, ANOVA, correlation, or regression, every result needs the same core elements. Missing any one of them will likely trigger a revision request.
| Element | Example | Why It Matters |
|---|---|---|
| Test statistic | t = 2.41, F = 5.73, r = .38 | Identifies which test was run and the magnitude of the result |
| Degrees of freedom | t(58), F(2, 87) | Tells the reader the sample size and test configuration |
| Exact p-value | p = .019 or p < .001 | Shows the probability of the result under the null hypothesis |
| Effect size | d = 0.63, η² = .12, r² = .14 | Shows practical significance beyond statistical significance |
| Confidence interval | 95% CI [0.28, 1.14] | Shows the range of plausible values for the effect |
| Group means and SDs | M = 72.4, SD = 9.1 | Gives the reader the actual scores, not just test outcomes |
If your analysis tool does not compute effect sizes automatically, you can calculate them by hand or run your data through Academic Stats Agent, which includes Cohen's d, η², and other effect sizes in every output. For a full explanation of when to use each effect size measure, see our effect size guide.
APA Formatting Rules for Statistical Results
APA 7th edition has specific formatting conventions for statistical reporting. These are small details, but reviewers and supervisors notice every one of them.
Italicize all statistical symbols: M, SD, t, F, p, N, n, d, r, χ², and η². Do not italicize numbers, parentheses, equal signs, or Greek letters used as variable names in equations. The rule applies in text and in tables.
Remove the leading zero from p-values. Write p = .034, not p = 0.034. The logic is that a p-value can never reach 1.00, so the zero before the decimal is unnecessary. The same rule applies to correlations (r = .42) and standardized regression coefficients (β = .31). Keep the leading zero on statistics that can exceed 1.00: M = 0.87, d = 0.55, t = 0.96. For a deeper look at what p-values mean and how to interpret them correctly, we have a separate guide.
Report exact p-values to two or three decimal places. Write p = .042, not p < .05. The only exception is when the p-value falls below .001. Then write p < .001 because three-decimal precision is sufficient and SPSS often rounds anything smaller to .000 anyway. Never write p = .000.
Use spaces around mathematical operators. Write t(42) = 2.38, p = .022 with spaces on both sides of the equals sign. Place degrees of freedom in parentheses directly after the test symbol with no space: F(2, 87), not F (2, 87).
How to Write a Results Paragraph
Each hypothesis should produce one clear paragraph. The structure is simple: state what you tested, name the test, give the numbers, say whether the hypothesis held. Here are two examples.
Example: Independent-Samples T-Test
Hypothesis 1 predicted that students who received peer tutoring would score higher on the final exam than students who did not. An independent-samples t-test was used to compare exam scores between the tutoring group (n = 34) and the control group (n = 31). Students in the tutoring group scored significantly higher (M = 76.8, SD = 10.2) than those in the control group (M = 70.3, SD = 11.5), t(63) = 2.41, p = .019, d = 0.60, 95% CI [1.11, 11.89]. Hypothesis 1 was supported.
Notice the pattern. Four sentences. No opinion about why tutoring helped. No citation of Vygotsky or zone of proximal development. Just data.
Example: One-Way ANOVA
Hypothesis 2 stated that job satisfaction would differ across the three management styles (authoritarian, democratic, laissez-faire). A one-way ANOVA was conducted with management style as the independent variable and job satisfaction score as the dependent variable. There was a statistically significant difference across the three groups, F(2, 117) = 6.84, p = .002, η² = .10. Post hoc comparisons using Tukey HSD indicated that the democratic group (M = 4.32, SD = 0.74) scored significantly higher than the authoritarian group (M = 3.71, SD = 0.89, p = .001), while the laissez-faire group (M = 4.05, SD = 0.81) did not differ significantly from either the democratic (p = .264) or authoritarian group (p = .112). Hypothesis 2 was partially supported.
The ANOVA paragraph is longer because it requires post hoc results. With three or more groups, you always need to tell the reader which specific pairs differed. An F-test alone only says "somewhere among these groups, there is a difference." For guidance on reporting SPSS output tables for these tests, see our SPSS output reporting guide.
Tables vs. Text: When to Use Each
A good rule: if you are comparing two groups, report the result in text. If you are comparing three or more groups, or if you have multiple dependent variables, use a table. A paragraph describing means and standard deviations for six groups across four variables would run to 300 words and be nearly unreadable. A table presents the same information in 15 rows.
Tables also work well for correlation matrices (when you have five or more variables), regression model summaries (with multiple predictors), and reliability coefficients across several subscales. In text, always refer to the table by number: "Table 3 presents the descriptive statistics for all study variables." Do not repeat every number from the table in the text. Pick the most important finding and highlight it; let the table carry the rest.
APA tables have no vertical lines and only three horizontal lines: one above the header row, one below it, and one at the bottom of the table. No gridlines, no shading. If your university has its own table format, follow that instead.
Common Mistakes (and How to Avoid Them)
After reviewing over 300 results chapters, we see the same errors come back repeatedly. Here are the ones that cause the most revision rounds.
Mixing results with discussion is the most frequent problem. A student writes "t(48) = 3.12, p = .003, which may be due to the fact that older participants had more work experience." That second clause is interpretation. Move it to the Discussion. The results section should end at "Hypothesis 1 was supported."
Writing "very significant" or "highly significant" is another common error. A result is either statistically significant or it is not. A p-value of .002 is not "more significant" than a p-value of .041. Both fall below the .05 threshold. If you want to communicate the strength of a finding, report the effect size. That is what effect sizes are for.
Reporting only significant findings is a form of selective reporting that weakens the entire chapter. If Hypothesis 3 was not supported, say so. Write the test, report the non-significant result (p = .348), and state that the hypothesis was not supported. Supervisors will ask about missing hypotheses, and reviewers for journal publication will flag it immediately.
Including raw data or SPSS screenshots in the body of the results section is another mistake. Raw data belongs in an appendix or a supplementary file. SPSS output belongs nowhere in the thesis body. Extract the numbers, format them in APA style, and write them into sentences. If your committee asks for the raw output, include it as an appendix and reference it: "The full SPSS output for all analyses is provided in Appendix C."
Forgetting effect sizes is the last frequent omission. APA 7th edition (Section 6.5) requires effect sizes for all primary results. A p-value tells you whether an effect exists; an effect size tells you how large it is. A study with N = 500 can produce a significant p-value for a difference of half a point on a 100-point scale. The effect size (d = 0.08) reveals that the difference, while statistically real, is practically meaningless.
Key takeaway: The results section reports findings without interpreting them. Structure it in three parts: preliminary analyses (descriptives, reliability, assumption checks), main analyses (hypothesis by hypothesis in order), and any exploratory analyses. For each test, report the test statistic, degrees of freedom, exact p-value, effect size, and group descriptives. Save the "why" for the Discussion chapter, report non-significant results honestly, and never paste a screenshot where a formatted sentence belongs.
Frequently Asked Questions
How do I report non-significant results in APA format?
Report them exactly like significant results: give the test statistic, degrees of freedom, exact p-value, and effect size, for example t(58) = 1.32, p = .192, d = 0.34. Never hide or omit non-significant findings; they are legitimate results that answer your research question.
Should I include exact p-values in my results section?
Yes. APA 7th edition requires exact p-values to two or three decimal places, such as p = .034, rather than p < .05. The only exception is very small values, which are reported as p < .001.
What is the difference between the results and discussion sections?
The results section reports what you found, with statistics, tables, and figures but no interpretation. The discussion section explains what those findings mean, how they relate to previous research, and what their limitations are.