What is effect size?

What is effect size?

• A. A numeric index of p-values
• B. The level of statistical significance
• C. A statistical expression of the magnitude of the difference between treatments or the magnitude of the relationship between variables
• D. The probability of making a Type I error

Answer: (C)

Effect size measures how big the observed difference is or how strong a relationship between variables appears. It tells you the magnitude of an effect, which is what matters for applying findings in practice, not just whether an effect exists. P-values tell you whether an effect might be due to chance, but they don’t tell you how large that effect is. That’s why effect size is essential: it communicates practical significance in clinical terms, independent of sample size. In physical therapy research you might see standardized mean differences like Cohen’s d for comparing two treatments, or correlation measures like r to describe the strength of a relationship, or odds ratios for binary outcomes—each conveys the real-world impact of the finding. The other ideas describe significance thresholds or error risk, not how large the effect is, so they don’t provide the magnitude information that effect size does.

Effect size measures how big the observed difference is or how strong a relationship between variables appears. It tells you the magnitude of an effect, which is what matters for applying findings in practice, not just whether an effect exists. P-values tell you whether an effect might be due to chance, but they don’t tell you how large that effect is. That’s why effect size is essential: it communicates practical significance in clinical terms, independent of sample size. In physical therapy research you might see standardized mean differences like Cohen’s d for comparing two treatments, or correlation measures like r to describe the strength of a relationship, or odds ratios for binary outcomes—each conveys the real-world impact of the finding. The other ideas describe significance thresholds or error risk, not how large the effect is, so they don’t provide the magnitude information that effect size does.