effektusméretek

Effektusméretek, often translated as effect sizes, are quantitative measures that describe the magnitude of a phenomenon, such as the difference between groups or the strength of a relationship. They provide information about practical significance and are often considered alongside p-values, since they aim to express how large an observed effect is rather than only whether it is statistically detectable.

Common effektusméretek include standardized mean differences (for example, Cohen’s d and Hedges’ g), which express the difference between two group means in standard deviation units; correlation coefficients (such as Pearson’s r and Spearman’s rho), which describe the strength of association between variables; and ratios such as odds ratio or risk ratio, which compare the odds or risks between groups. Other measures estimate the proportion of variance explained, such as eta-squared and omega-squared, and there are also various more specialized indices like partial eta-squared and nonparametric or distribution-free measures.

Interpreting effektusméretek depends on the context and the chosen measure. Some rules of thumb exist—for example, small, medium, and large categories for Cohen’s d or for Pearson’s r—but thresholds vary by field and study design. Effektusméretek are widely used in planning studies (power analysis), reporting results, and conducting meta-analyses, where they facilitate combining findings across studies.

Limitations include that different measures are not directly comparable, they can be sensitive to study design and measurement scales, and misinterpretation is common when generalizing across contexts. Reporting guidelines often recommend including the estimate, its confidence interval, and a clear description of the measure used.