Results-chapter generator: from calculator outputs

Upload the output files

Arranging the analyses by hypothesis

What the generator does – and what it does not

Every calculator on the site produces, for each analysis, tables in APA format, table notes, and report sentences in English. Someone who runs several analyses – a t-test for three variables, correlations, regression – gets several separate files. The generator collects them into one chapter: it reads the HTML files the calculators export, identifies in each file the calculator, the data source and the analyses (by the headings), and builds from them a results chapter in the order you set – with the report sentences as the calculator worded them, the tables renumbered in a single sequence, the notes, and an opening paragraph about the data sources and the sample size. The generator computes nothing and words nothing new: every number and every sentence comes from the calculator that produced it; what the generator adds is order, numbering, headings and a checklist. The site convention: there is no artificial intelligence here and no automatic wording – the wording of the hypotheses is yours, and the chapter you download is a draft for editing, not a final document.

What does not go into the chapter. The calculators' "Interpreting the result" box (the interpretation, the caveats, the references to the books) is not part of the results chapter – according to APA (2020, section 3.7) the results chapter reports, and interpretation belongs in the discussion (section 3.8). It can be attached as an appendix "Notes for the discussion" (a checkbox), so that the calculator's notes – for example on a confidence interval that includes zero or on a difference between the Welch row and the equal-variances row – serve you when writing the discussion. Charts are not in the HTML file the calculator exports (they stay on the calculator page) – figures are attached separately, according to the figure rules on the guidelines page. And outputs of the preparation pages (normality testing, the standard template) suit the method section, not the results chapter – see "What belongs in the method section".

The structure of the results chapter according to APA

APA (2020, section 3.7) defines what the results chapter of a quantitative study should contain, and in what order. The order – the same structure as in the method section: primary hypotheses, then secondary, and then exploratory analyses; this is why the generator asks for the type of hypothesis for each analysis and orders by it (whoever prefers a different order chooses "In list order"). Completeness – "report all relevant results, regardless of whether your hypotheses were supported, including results that run counter to expectation"; a non-significant result is no reason to omit an analysis. The minimum for every test – the value of the statistic, the degrees of freedom, an exact p-value, and the effect size with its direction; for a point estimate (mean, coefficient) – a measure of precision (standard error or confidence interval); and the "sufficient set" of statistics – the size of each cell, the cell means (or frequencies for a categorical variable) and the standard deviations – is what allows the reader to verify the result and include it in a meta-analysis. Confidence intervals – "generally the best reporting strategy", at a single confidence level throughout the paper, set in advance (95% or 99%); and above all – "reporting all tested hypotheses, estimates of effect sizes, and confidence intervals are the minimum expectations for all APA journals" (Wilkinson and the Task Force, 1999, as cited in APA). Ancillary and follow-up analyses – subgroup analyses and adjusted analyses are reported too, distinguishing between what was planned in advance and what is exploratory, and with a discussion of the implication for error rates. The calculators' report sentences are built so that all of this is in them; the reporting guidelines page details what must be in the sentence for each test.

How it works: from the calculator to the chapter

1. In the calculator. After running, in the export row at the bottom of the output – "Download as HTML". The file contains everything in the output except the charts and the copy buttons: tables, notes, report sentences and the interpretation of the result. Run every analysis you want in the chapter this way; if one test was run on several dependent variables – one file is enough, each variable is a separate analysis.

2. Here. Drag in the files (several at once is possible, and more can be added later). The generator shows a row for each analysis: the calculator, the analysis name (for example "age" in a t-test), the data source and N. For each row, choose a hypothesis type, write the wording of the hypothesis (optional – for example "H1: the experimental group will report more positive attitudes"), and check whether to include the tables. The arrows move an analysis up or down; an analysis that is not needed – remove it.

3. The chapter. "Build the chapter" produces the chapter in the output box: a title, an opening paragraph (data sources, N, how many analyses), and then a subheading for each group of hypotheses and for each analysis – the wording of the hypothesis, the report sentence (in the chosen language), the tables with their notes. At the end – the checklist. Download as Word (each table – a real Word table) or as HTML, and edit: add the context, delete surplus tables, and decide what stays in the text and what moves to a table (APA, 2020, section 7.3: the result of one test or a few means – in a sentence; a table – for the results of many tests or for extensive descriptive statistics).

The automatic checklist

At the end of the chapter the generator attaches a checking table: for each analysis – whether the report sentence and its tables contain (a) a statistic value with degrees of freedom (t, F, χ², U, z, H, r…), (b) an exact p-value or "< .001", (c) an effect size measure (d, η², ω², r, r_rb, R², β, φ, V, OR, κ, α…), (d) a confidence interval, and (e) the sample size. The check is textual – it looks for the symbols in the text the calculator produced – so it is a reminder list, not a seal of approval: ✓ means the component appears, ✗ means it was not found and that it is worth checking whether it is required. In most of the site's tests everything appears; in descriptive measures and in reliability some components are not relevant (there is no p and no effect size for a descriptive measure) – and the table marks this. Below the table – two questions the generator cannot answer for you, and which according to APA (2020, section 3.7) are a condition for the chapter: whether the handling of missing data was reported (the rate of missing data, the presumed reason, the handling method) and whether follow-up analyses were distinguished from the planned ones.

Tables, numbering and references

The calculators number tables from 1 in each output; in a single chapter the numbering must be consecutive – so the generator renumbers in order of appearance (APA, 2020, section 7.10: tables are numbered in the order in which they are mentioned in the text). The titles are kept as the calculator worded them (short, in italics, below the number – section 7.11), and so are the table notes (section 7.14: a general note, then specific, then probabilities). Note two things: references in the text – the calculators' report sentences do not refer to a table ("see Table 3"), because they were written without knowing the final number; after downloading, add the reference in the sentence, according to the number in the chapter. Duplication – APA (section 7.3) warns that a table that repeats the text may be unnecessary; the calculators give both on purpose (the sentence for the text, the table for an appendix or for the text), and you choose when editing. Whoever wants a chapter of sentences only unchecks "Tables" in the relevant rows.

What belongs in the method section

APA (2020, sections 3.6-3.7) asks that data screening – outliers, assumptions, distributions – be reported in the method section, before the results, together with the missing data ("the frequency or percentages of missing data and the methods for addressing it"), and that the decisions derived from it (exclusions, transformations, imputations) be mentioned. Therefore the output of normality testing or the check report of the standard template is not material for the results chapter but for the method section; the generator will accept them (they are valid HTML files of the site) but will mark them as "Screening – for the method section" and place them at the opening of the chapter, before the hypotheses, so that you move them to their place. Reliability (α) too is reported, according to Pallant (2020, chapter 9), in the method section under "Measures" – the generator marks it in the same way. The assumption tests within the test output – Levene, normality of residuals, VIF – stay in their tables; when editing, decide whether they stay in the results or move up to the method section.

What belongs in the discussion

According to APA (2020, section 3.8) the discussion opens with "a clear statement of the support or nonsupport for all hypotheses, distinguished by primary and secondary hypotheses"; in the case of an ambiguous result – explain why it is considered so; exploratory analyses are also discussed in terms of the uncontrolled error rates. After that – similarities and differences with other studies, limitations and strengths (biases, imprecision of measures, the number of tests, sample size and sampling validity), and implications. The generator does not write a discussion – but two things in it help write one: the order of the hypotheses (the discussion goes through them in the same order), and the appendix "Notes for the discussion" – the calculators' "Interpreting the result" boxes, which already state for each analysis what p, the effect size and the confidence interval mean, with the references to the books. The abstract (sections 2.9 and 3.3), the opening of the discussion (the statement of support for each hypothesis) and the executive summary for the client are built from the form of the second generator – Results-chapter and executive-summary generator: from free input; the discussion itself you write.

Common mistakes

1. Omitting an analysis that did not come out significant. APA (section 3.7): "do not hide uncomfortable results by omission". A planned analysis is reported, even when p ≥ .05 – and with the effect size and the confidence interval, which say whether there was a small effect here or a small sample.
2. Presenting exploratory as planned. What was not planned in advance is marked as exploratory – in the results chapter and in the discussion. Changing the classification after seeing the results is exactly what the APA order is meant to prevent.
3. Reporting the same number twice. A table that repeats what is already in the sentence is unnecessary (section 7.3) – in the sentence or in the table. The generator gives both so that you choose.
4. Forgetting the reference to the table. Every table is mentioned in the text by its number, and the reader is told what to look for in it ("As shown in Table 2…"; section 7.5 – not "the table below") – the final number is known only after assembly, so the reference is added when editing.
5. Mixing interpretation into the results. "Hence the program succeeded" – for the discussion. In the results: what was found, how much, and with what certainty.
6. Leaving the screening in the results. Normality, outliers, missing data – for the method section (section 3.6), before the results.
7. Shifting confidence levels. 95% in one analysis and 99% in another – APA asks for one level, set in advance, for the whole paper.
8. Trimming the "sufficient set". Without N, means and standard deviations for every cell it is impossible to verify and include in a meta-analysis – the calculators' tables contain them; if you delete a table, these numbers must appear in the sentence.

References

American Psychological Association. (2020). Publication manual of the American Psychological Association (7th ed.). https://doi.org/10.1037/0000165-000

Pallant, J. (2020). SPSS survival manual: A step by step guide to data analysis using IBM SPSS (7th ed.). Routledge.

Wilkinson and the Task Force on Statistical Inference (1999) – as cited in APA (2020, section 3.7). Per the site rule, the reference list includes only sources that were read in the original.