The pooled standard deviation combines the spreads of several groups into one estimate, weighting each group by its degrees of freedom. It is used in Student’s t-test, ANOVA and Cohen’s d.
sₚ = √[ Σ(nᵢ − 1)sᵢ² ÷ Σ(nᵢ − 1) ]
Pooling assumes the groups share the same population variance. If their SDs differ a lot, especially with unequal sizes, prefer Welch’s methods, which don’t pool.
Example
Two groups: s₁ = 8.5 with n₁ = 20 and s₂ = 9.2 with n₂ = 18. sₚ = √((19 × 72.25 + 17 × 84.64) ÷ 36) = √78.10 = 8.84. Cohen’s d for a 6-point mean difference is then 6 ÷ 8.84 = 0.68.