Test whether two populations have equal variances. Essential for checking assumptions before pooled t-tests, and for assessing process consistency and stability.
F-statistic (numerator has the larger variance):
F = s₁² / s₂²
df₁ = n₁ − 1, df₂ = n₂ − 1
Two-tailed p-value:
p = 2 × P(F(df₁, df₂) > F_observed) if F > 1
CI for variance ratio (σ₁²/σ₂²):
[ F / F_upper, F / F_lower ]
where F_upper and F_lower are critical F values at α/2.
Note: F-test is sensitive to non-normality. For robustness, consider Levene's test.