Medical researchers commonly report the effects of a treatment or describe relationships between variables. Treatment effects or associations can be quantified using measures like mean differences, risk ratios, or correlations. Null hypothesis (Ho) significance tests (eg, t tests or χ2 tests) are commonly used to determine the “statistical significance” of the observed effect, usually defined by a P value of <.05. However, P values are often misinterpreted and provide no information on the magnitude or importance of the effect.1–3 Hence, rather than merely focusing on statistical significance, researchers should provide plausible estimates about the magnitude of the effect in the population from which the data were sampled.4,5 Previous statistical tutorials in this series initially discussed P values, confidence intervals (CIs), and effect size.6–8.

The aim of the present basic statistical tutorial is to discuss in greater detail how a treatment effect or association can be quantified using the effect size, and how a CI can help to assess the statistical but especially also the clinical significance of the observed effect. Moreover, we discuss what P values actually represent and how they should be interpreted. We will recurrently use a study by Frey et al,9 which analyzed perioperative temperature management in 79 patients undergoing open colon surgery, to illustrate these concepts. These study patients were randomly assigned to standard temperature management, or standard management in combination with insufflation of warm, humidified carbon dioxide into the wound cavity. While these authors report several outcomes, we specifically focus here on the difference in core temperature at the end of surgery

This article is intended to help students cement their understanding of many concepts including “statistical significance”. Explain your understanding of this concept  in a paragraph form.