Cohen's d is a standardized effect size for continuous outcomes, expressing the difference between two means in units of the outcome's standard deviation.
Why standardization helps
Different trials measure focus, mood, sleep quality, or soreness on different scales. Cohen's d puts the between-group difference onto a common unit so results are easier to compare across instruments.
The metric is most useful when the outcome is continuous and the groups are comparable. It is less natural for time-to-event outcomes, binary outcomes, or outcomes with very skewed distributions.
How it is calculated
In a simple two-group trial, Cohen's d is the mean difference divided by a pooled standard deviation. A positive or negative sign shows direction, so the interpretation depends on whether higher scores are better or worse.
Small samples can make d unstable. That is why the point estimate should travel with a confidence interval, especially when a study is used to support an evidence tier.
What it can and cannot say
Cohen's d says how large the average separation was relative to variation in the sample. It does not say how many people responded, whether the change was clinically meaningful, or whether the same size applies to a different population.
In supplement research, the same d value may feel different depending on the outcome. A modest shift in sleep latency may matter more to a user than the same standardized shift in a questionnaire subscale.
Use in pooled evidence
A meta-analysis may pool standardized mean differences when studies report related outcomes on different scales. This can help summarize a field, though it can also hide differences in measurement quality.
For recommendations, Cohen's d is useful only when paired with study design, population match, dose match, and uncertainty.
Safety note
Cohen's d is an efficacy metric. It does not measure adverse-event risk, interaction risk, or the personal cost of trying an intervention.