Statistical Concepts and Testing

Statistical Concepts and Testing

Statistical testing determines whether observed differences or associations in data are likely to reflect real effects or could plausibly be due to chance. For emergency medicine trainees (FRCEM) and clinicians, understanding the purpose, assumptions and limitations of common tests - and how to interpret effect sizes and confidence intervals alongside p‑values - is essential for exams and for critically appraising the literature.

This section covers core definitions, a practical guide to choosing common tests, interpretation (statistical vs clinical significance), reporting expectations aligned with UK guidance, common pitfalls, and brief worked examples.

Core concepts - concise definitions

• Null hypothesis (H0): the default assumption that there is no difference, effect or association between groups or variables. Statistical tests assess how compatible the data are with H0.
• P‑value: the probability of observing data at least as extreme as those obtained, assuming H0 is true. A conventional threshold is p < 0.05 for “statistical significance,” but p‑values do not measure effect size or clinical importance and are sample‑size dependent.
• 95% confidence interval (CI): a range that, under repeated sampling, would contain the true population parameter 95% of the time. CIs show both the estimate and its precision. For differences, CIs not crossing 0 imply statistical significance; for ratios (RR, OR, HR), CIs not crossing 1 imply statistical significance.
• Effect size:...