What belongs in a statistical results chapter?

Give the reader a route through the findings

Organize results around the research questions or prespecified comparisons rather than a dump of software output. For each question, identify the analyzed participants or units, describe the variables, show the relevant estimate and explain its uncertainty and limitations. Your university or journal may require a particular order or format; follow those instructions first.

The APA Journal Article Reporting Standards describe what quantitative papers should report across their sections. The SAMPL reporting guidance provides more detailed statistical reporting recommendations. These are useful checks for completeness, not a template that overrides a supervisor's requirements.

What a reader should be able to find

• Who was analyzed? Give the relevant sample size for each analysis, not just the number initially recruited. Explain exclusions and missing observations when they matter.
• What was measured? Define scores, groups, time points and units. A mean without a scale or group label is hard to interpret.
• What analysis answered the question? Name the approach and report the key assumptions or diagnostics that affect interpretation.
• What is the estimated result? Present an effect, difference or association with a suitable measure of uncertainty when possible. A p value alone does not tell the reader the size or practical meaning of an effect.
• What are the limits? Sampling, measurement, missing data and study design can narrow the conclusion. Observational association is not by itself proof of causation.

A worked example from a complete study

The academic self-efficacy and persistence example contains 180 students, two five-item scales and one association hypothesis. Both scales use mean scores on a 1–5 response range; this dataset has no missing responses or reverse-scored items.

For what these coefficients can and cannot establish, see the guide to Cronbach’s alpha and questionnaire reliability.

A reporting paragraph could read: “Academic self-efficacy was positively associated with learning persistence, r = .43, 95% CI [.31, .55], p < .001, using a two-sided Pearson test with 180 participants. Students reporting higher self-efficacy tended also to report higher persistence.”

The estimate describes the strength and direction of the association; the interval conveys its uncertainty. A sentence such as “self-efficacy improves persistence” adds a causal claim that this cross-sectional design cannot establish. The mean scores and reliability estimates provide context, rather than replacing the hypothesis result.

Open the full results chapter, prepared data and software output to follow the numbers back to their sources. In the advanced example, preliminary factor analysis belongs to measurement assessment, while the explicitly additional model is reported separately from the main hypotheses.

Tables, text and software files have different jobs

A table lets readers compare numbers. The accompanying text highlights the result needed to answer the research question; it need not repeat every cell. An output file can help reproduce or inspect a calculation, but it is not a polished explanation. If a package includes SPSS-compatible files, compatibility alone does not establish that IBM SPSS performed the computation.

Before handing in the chapter

Check that the same sample, variable labels, estimates and rounding appear in the methods, tables and prose. Report non-significant results rather than deleting them because they were unexpected. Check that the claims in your discussion stay within what the design and results support. The complete research examples connect study descriptions, analysis plans and downloadable outputs, and the deliverables page explains how a scoped results package is defined.

For a focused worked table, explicit denominators and guidance on choosing center and spread, read the descriptive statistics guide.

For an interpretation of B, standardized coefficients, confidence intervals and the limits of comparing an added model, read the multiple linear regression guide.

See the complete research examples

Follow the research questions through the input files, analysis plan and outputs. Each example includes the consultation report, itemized illustrative price and generated files.