A quantitative results chapter follows a fixed order: data screening, sample description, measurement quality (reliability such as Cronbach's alpha ≥ .70, Nunnally, 1978, and validity), then hypothesis tests reported with test statistic, degrees of freedom, exact p value, effect size and confidence interval, as APA 7 requires (American Psychological Association, 2020). It reports what you found without interpreting why; interpretation belongs in the discussion chapter. Each hypothesis should end with a clear "supported" or "not supported".
The standard structure
- Introduction (one paragraph): what the chapter covers and in what order.
- Data screening: responses collected versus usable; how you handled missing data, straight-lining and failed attention checks; outliers; reverse coding.
- Sample profile: a demographic table with frequencies and percentages.
- Descriptive statistics: means, standard deviations, and the correlation matrix of the main variables.
- Measurement quality: reliability (alpha, composite reliability), EFA or CFA, convergent and discriminant validity (AVE ≥ .50, Fornell & Larcker, 1981; HTMT < .85, Henseler et al., 2015).
- Assumption checks for your main test: normality where it matters, multicollinearity (VIF), homoscedasticity.
- Hypothesis tests, in the order of your hypotheses.
- Summary table of hypotheses and decisions.
Rules that examiners check
- Every statistic complete: for example t(98) = 2.45, p = .016, d = 0.49, not "p < .05" alone. APA 7 asks for exact p values to two or three decimals, with p < .001 for very small values (APA, 2020).
- Effect sizes with a cited benchmark, such as Cohen's (1988) .02, .15 and .35 for f².
- No leading zero for statistics that cannot exceed 1 (p, r, α, β); keep it for M, SD, d and t.
- Tables and text work together: do not repeat every number of a table in the text; point to the table and highlight what matters.
- Every threshold sourced: "acceptable" must always mean "above X (Author, Year)".
- No interpretation beyond what the numbers say. Save "this suggests that…" for the discussion.
Worked example: one hypothesis, written up
H1 states that perceived workload negatively predicts job satisfaction. The student ran a multiple regression with workload, autonomy and supervisor support as predictors, N = 112. The model was significant, F(3, 108) = 20.25, p < .001, R² = .36, adjusted R² = .34. Using f² = R² / (1 − R²) = .36 / .64 = .56, the model effect is large by Cohen's (1988) benchmark of .35. VIF values were between 1.18 and 1.52, well below 5 (Hair et al., 2019). The paragraph:
A multiple regression was conducted to test H1–H3. Multicollinearity was not a concern (VIF ≤ 1.52; Hair et al., 2019). The model explained 36% of the variance in job satisfaction, F(3, 108) = 20.25, p < .001, R² = .36, f² = .56, a large effect (Cohen, 1988). Workload had a significant negative effect on job satisfaction, β = −.31, t(108) = −3.87, p < .001, 95% CI [−.47, −.15]. H1 was therefore supported. Coefficients for all predictors are shown in Table 5.
Check the degrees of freedom: with three predictors and 112 cases, the regression has 3 and 112 − 3 − 1 = 108 degrees of freedom, and each coefficient's t test has 108. Errors like this are easy for an examiner to spot.
Tables to include
- Sample profile (demographics).
- Descriptives and correlations with reliability on the diagonal.
- Measurement model: loadings, alpha, composite reliability, AVE.
- Discriminant validity: HTMT or Fornell-Larcker matrix.
- Hypothesis tests: coefficients, t, p, CI, effect sizes.
- Summary of hypotheses with decisions.
Format each in APA 7: bold table number, italic title, horizontal rules only, and a note giving N, abbreviations and threshold sources (APA, 2020).
Non-significant results
Report them exactly like significant ones, with the full statistics, and state "not supported". Do not drop them or call them "approaching significance". If your sample was small, a sensitivity power analysis helps the discussion chapter.
Run it free
Paste a draft into the free results doctor, which checks each paragraph for missing statistics, wrong decimals and unsourced thresholds. Not sure which test answers your hypothesis? Use the statistical test chooser.
FAQ
How long should the results chapter be?
Long enough to report every analysis your method chapter promised. In master's dissertations it is often the shortest of the main chapters because interpretation is left out.
Should I put SPSS output in the chapter?
No. Rebuild the tables in APA format. Raw output can go in an appendix if required.
Do I discuss literature in the results?
Not usually. Comparison with prior studies belongs in the discussion, unless your department combines the two chapters.
One-tailed or two-tailed tests?
Two-tailed unless you pre-specified directional hypotheses and say so.
References
- American Psychological Association. (2020). Publication manual of the American Psychological Association (7th ed.). https://doi.org/10.1037/0000165-000
- Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum Associates.
- Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. https://doi.org/10.1177/002224378101800104
- Hair, J. F., Risher, J. J., Sarstedt, M., & Ringle, C. M. (2019). When to use and how to report the results of PLS-SEM. European Business Review, 31(1), 2–24. https://doi.org/10.1108/EBR-11-2018-0203
- Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science, 43(1), 115–135. https://doi.org/10.1007/s11747-014-0403-8
- Nunnally, J. C. (1978). Psychometric theory (2nd ed.). McGraw-Hill.