Chapter 17. Writing the Results Chapter
17.1 Purpose and boundaries
The results chapter reports what the analysis found. It should answer the research questions in a logical order, explain exclusions and missing data, describe the sample, present primary results, and report relevant supplementary or sensitivity analyses. Save the full theoretical explanation and comparison with prior studies for the discussion chapter, although brief neutral signposting may be appropriate.
17.2 Recommended order
- Chapter introduction and reminder of the analytical purpose.
- Data screening, exclusions, missingness and final analytic sample.
- Participant or organisational characteristics.
- Measurement and reliability/validity evidence, where relevant.
- Descriptive statistics for key variables.
- Results for each research question or hypothesis in the same order as the objectives.
- Model diagnostics and sensitivity analyses.
- Short summary of the most important findings without over-interpreting them.
17.3 Reporting numbers
Use consistent decimal places and units. Include the denominator for percentages where ambiguity is possible. Report exact p-values to sensible precision (for example, p = .032); for very small values, report p < .001 rather than p = .000. Report effect estimates and confidence intervals. State the analytic sample size for each model if missing data cause it to vary. Follow the required style guide for statistical notation.
17.4 Table and figure design
Every table and figure should have a number, concise title, clear labels and explanatory notes. Avoid repeating every table value in prose; instead, draw attention to the result that answers the question. A regression table should identify the outcome, predictors, reference categories, coefficient type, uncertainty, sample size and model fit information relevant to the model. Never paste unexplained software screenshots as the final presentation of results.
17.5 Results reporting example
Template: “The analytic sample included [n] participants. The mean [outcome] score was [M] (SD = [SD]; 95% CI [L, U]). In the prespecified regression model, [predictor] was associated with [outcome] (B = [value], 95% CI [L, U], p = [value]), adjusting for [covariates]. The estimate indicates [plain-language magnitude]. The model explained [appropriate measure] of variation in the outcome.” This is a template, not a finding; insert only verified results.