Which statistic quantifies effect size when continuous outcomes are on different scales?

• A. Odds ratio
• B. Cohen's d
• C. Risk difference
• D. Hazard ratio

Answer: (B)

When continuous outcomes are measured on different scales, you need a standardized, unitless estimate of how large the difference is. Cohen's d provides that by taking the difference between group means and dividing it by the pooled standard deviation. This yields a dimensionless effect size, so you can compare effects across studies or measures that use different units. It also conveys the magnitude of the difference relative to variability, which is what you want when scales differ.

The other options are intended for different types of data: an odds ratio, risk difference, and hazard ratio all pertain to binary outcomes or time-to-event data, not to continuous outcomes on differing scales. They don’t standardize a difference in means across scales, so they aren’t appropriate for this scenario.