The results section looks the easiest to write because you “only” have to put in the numbers. And it is the one that gathers the most errors: figures that do not match the tables, conclusions that slip in early, tests reported without their statistic, results told in an order the reader cannot follow.
Its function is a single one: to report what you found, completely and in order, without explaining it yet. The explanation belongs in the discussion. Mixing them is the commonest mistake, and it makes the results read like an opinion and the discussion like a repetition.
This guide gives you the order of the section, the rule for choosing between text and table, an example with figures that agree and the list of errors reviewers see most. For the exact APA 7 format of each statistic, see also the blog articles linked below.
It reports what you found in the order the reader needs to understand it, with complete statistics so they can judge it. It does not justify why it came out that way, compare it with other studies or draw implications: that is the discussion. A good test is whether someone can read only your results and reconstruct what happened in the study, even if they do not know what you think about it.
A practical nuance: reporting is not the same as interpreting in other words. Phrases such as “which indicates that the intervention was effective” are interpretation and go to the discussion. In results, “the intervention group scored lower than the waitlist group” is enough.
An order that serves almost any study follows the logic from general to particular:
Give each hypothesis a paragraph, or a subsection. Open with a sentence that says what was tested and end with the figure that answers it. That way the reader knows at every moment which question each number belongs to, and you check along the way that every hypothesis has an answer and that there are no orphan results.
The practical rule is that the results that carry the story go in the text, and the details a reader might want to check go in tables. With three numbers or fewer, write in the text. With more, use a table and comment on the important part in the text: what is compared and what the pattern is, without repeating each cell.
A table transcribed in full into the text is a classic error: it takes space, adds nothing and raises the chance that a figure does not match. A figure beats a table when what matters is a pattern (an interaction, a trend, a distribution) more than the exact values. And every table or figure is cited in the text and stands on its own with its title and note.
“The intervention group (n = 15) scored lower on anxiety after the programme (M = 14.2, SD = 3.4) than the waitlist group (n = 15, M = 16.9, SD = 3.6), t(28) = 2.11, p = .044, d = 0.77, 95% CI of the difference [0.08, 5.32].”
In one sentence there is everything a reader needs to check it: the groups and their size, the means and standard deviations, the statistic with its degrees of freedom, the exact p value, the effect size and the interval. And the figures agree: a mean difference of 2.7 with those standard deviations and 28 degrees of freedom gives that t, that p and that d. It is exactly what a reviewer recomputes, which is why they should match.
For the non-significant result the format is the same and it is reported just as fully: “At three months the difference was not significant, t(26) = 1.10, p = .281, d = 0.42”, with its confidence interval alongside. Do not hide it or replace it with “there were no differences”: the interval says which values it still allows.
A reviewer finds them with a quick read and a calculator:
Do one pass with the original output of the analyses beside you and check, number by number, that every figure in the text and tables matches it and the abstract, that each p value corresponds to its statistic and that each hypothesis has its answer. It is the dullest pass and the one that finds the most errors, and it can be partly automated.
For the exact format of each statistic in APA 7 (when to italicise, how many decimals, how to report p < .001), see the blog articles on how to report results in APA 7 and the full example of a results section commented sentence by sentence.
Paste your results section and you will see the first objection a reviewer would raise, with the exact sentence from your text and what it would take to fix it. It checks what this guide goes through: whether every test comes with its statistic, degrees of freedom, p value and effect size with an interval, whether assumptions are checked, whether multiple comparisons go unadjusted, whether the numbers in the text match the tables, and whether you interpret where you should only report.
It returns one objection, the one that weighs most. It needs at least 350 characters and what you paste is not stored. If you also want every p value recomputed from its test statistic, the guide on p values that do not match runs that check, and Previum does both on the whole manuscript.
Results report what was found, without explaining it; the discussion interprets those results, compares them with the literature and draws implications. In results you do not opine, and in the discussion you do not repeat the figures more than needed.
With three or fewer, write in the text; with more, use a table and comment on what matters. Transcribing a whole table into the text takes space and raises the risk of a figure not matching.
Yes, with the same format and detail as the significant ones. Omitting them or replacing them with “there were no differences” is reporting bias, and a reviewer who knows the hypotheses will miss them.
Better not to. A p of .06 is not significant at a .05 threshold, and calling it a “trend” or “marginal” is a way of interpreting it in favour of the hypothesis. Report the value, the effect size and the interval, and let the reader judge.
From general to particular: the sample, the descriptives, the main analysis hypothesis by hypothesis, the secondary analyses and the exploratory ones, clearly separated.
The exact value to two or three decimals and without a leading zero (p = .044), except for very small values, written p < .001. The p is italic and is reported together with the statistic and its degrees of freedom.