Effect size

An effect size is a quantitative measure of the magnitude of a difference, association, or treatment effect. It can be expressed in original units or on a standardized scale.

Statistical significance: A judgment about whether observed data are sufficiently inconsistent with a specified null hypothesis. A small effect can be statistically significant with enough data, while a large estimate can remain uncertain.
Cohen's d: A standardized mean difference obtained by dividing a difference between two means by a standard deviation. It expresses a group difference in standard-deviation units.
Meta-analysis: A statistical method for combining results from multiple studies addressing a related question. Meta-analyses combine effect estimates to summarize evidence across studies.
P-value: The probability, under a specified null hypothesis and model, of obtaining a test statistic at least as extreme as observed. Unlike an effect size, it does not directly quantify magnitude.
Confidence interval: A range produced by a statistical procedure that captures a parameter across repeated samples at a specified rate. Intervals show the uncertainty around an estimated effect size.
Hedges' g: A bias-corrected standardized mean difference commonly used for small-sample studies. It adjusts Cohen's d for small-sample upward bias.
Sample size determination: The process of choosing the number of observations needed for a study's planned analysis. Researchers specify a target effect size when planning adequate sample sizes.
Relative risk: The ratio of outcome probabilities in two groups. It describes relative change, unlike the absolute change measured by risk difference.
Statistical power: The probability that a statistical test rejects its null hypothesis when a specified alternative is true. Power depends partly on the effect size a study is designed to detect.
Pearson correlation coefficient: A statistic measuring the direction and strength of a linear association between two quantitative variables. It serves as an effect size for linear association.

Linked from 20 pages