Which statement about effect size measures such as Cohen's d, Glass's g, and r is true?

Which statement about effect size measures such as Cohen's d, Glass's g, and r is true?

• A. They quantify the magnitude of an observed effect independent of sample size
• B. They are p-values
• C. They must be greater than 0.05 to be meaningful
• D. They replace the mean difference

Answer: (A)

Effect size measures quantify how large an observed effect is, independent of sample size. Cohen's d represents the standardized difference between two means, typically by dividing the mean difference by a pooled standard deviation, giving a unitless value that shows how many standard deviations apart the groups are. Glass's g is similar but uses the standard deviation of the control group, which can be helpful when variances differ between groups. The correlation coefficient r expresses the strength and direction of the association between two variables, with larger absolute values indicating stronger relationships. The important point is that these statistics convey magnitude rather than significance—they tell you how big the effect is. Significance tests produce p-values that depend on sample size, so a small effect can be significant with a large sample and a large effect can be nonsignificant with a tiny sample. Effect sizes do not replace the raw mean difference; they complement it by providing a scale-free sense of magnitude across studies.

Effect size measures quantify how large an observed effect is, independent of sample size. Cohen's d represents the standardized difference between two means, typically by dividing the mean difference by a pooled standard deviation, giving a unitless value that shows how many standard deviations apart the groups are. Glass's g is similar but uses the standard deviation of the control group, which can be helpful when variances differ between groups. The correlation coefficient r expresses the strength and direction of the association between two variables, with larger absolute values indicating stronger relationships. The important point is that these statistics convey magnitude rather than significance—they tell you how big the effect is. Significance tests produce p-values that depend on sample size, so a small effect can be significant with a large sample and a large effect can be nonsignificant with a tiny sample. Effect sizes do not replace the raw mean difference; they complement it by providing a scale-free sense of magnitude across studies.