Effect size
Effect size is a statistical measure of how large a difference, association, or intervention effect is.
What effect size is
Effect size is a way to describe how large a finding is. A study might estimate the difference between two group means, the strength of a correlation, the change after an intervention, or the relative risk of an outcome. The exact measure depends on the research question and the data.
Statistical versus practical meaning
A result can be statistically significant but too small to matter in practice, especially in a very large sample. It can also be practically important but uncertain in a small or noisy study. Effect size helps separate the size of the observed effect from the question of whether the evidence is strong enough to rule out chance.
Raw and standardized measures
Some effect sizes stay in the original units, such as a five-point change on a symptom scale or a two-kilogram difference in weight. Standardized effect sizes divide by a scale such as the standard deviation, making results easier to compare when studies use different measurement scales. Cohen's d is a well-known standardized mean difference.
Associations and risks
Effect size is not limited to differences between averages. A correlation describes the strength and direction of association between two variables. An odds ratio, risk ratio, or risk difference describes how much an outcome differs between groups. Each answers a different kind of magnitude question.
Confidence intervals
An effect size is an estimate, so it should usually be read with uncertainty. A confidence interval shows a range of values compatible with the data under the statistical model. Wide intervals warn that the estimate is imprecise; narrow intervals make it easier to judge whether the effect is large enough to matter.
Role in power analysis
Power analysis asks how likely a study design is to detect a specified effect. Larger expected effects usually require smaller samples to detect, while smaller meaningful effects require more data. This is why planning a study often begins by deciding what effect size would be scientifically, clinically, or practically meaningful.
Meta-analysis
Meta-analyses often combine effect sizes from multiple studies. When studies use different scales, standardized measures can put results onto a common metric. The combined estimate still depends on study quality, publication bias, heterogeneity, and how uncertainty is weighted across studies.
Why it matters
Effect size keeps attention on magnitude. It helps readers ask whether a treatment changes outcomes enough to justify cost or risk, whether a relationship is strong enough to guide decisions, and whether a statistically detectable pattern is meaningful outside the dataset.