F 方差检验计算器

F = s₁² / s₂²

Enter two sample variances and their degrees of freedom to compute the F statistic and critical value and test whether two population variances are equal.

📐 计算公式
双样本方差 F 检验:F = s₁² ÷ s₂²,自由度 df₁ = n₁−1、df₂ = n₂−1;右尾/左尾概率由不完全 Beta 函数 betai 数值求出,双侧 p = 2×min(右尾,左尾),输出截到 ≤1。校验:s₁、s₂、n₁、n₂ 四项非空守卫,结果表另要求方差 >0 且样本量 ≥2。判据:p <0.05 判「方差不齐(差异显著)」,否则「未见方差不齐」;附 α=0.05 双侧 F 临界值对照表(SD 比 = √F)。

检验两总体方差是否相等。

Formula & Standard References

Item Formula / Note Source / Standard
F statistic F = s1^2 / s2^2 (larger variance conventionally placed in the numerator) Fisher RA, Statistical Methods for Research Workers (1925)
Degrees of freedom df1 = n1 - 1, df2 = n2 - 1 NIST/SEMATECH e-Handbook of Statistical Methods §1.3.5.9
Critical value Reject when F > F(1 - alpha; df1, df2) for a right-tailed test NIST/SEMATECH e-Handbook of Statistical Methods, F distribution
Assumptions Both populations approximately normal and samples independent; not robust to non-normality NIST/SEMATECH e-Handbook of Statistical Methods §1.3.5.9
📌 计算说明
  • 通常取较大方差作分子。
  • 双侧 p = 2·min(右尾,左尾)。

📚 深度解析:F 检验(方差比)

💡 常见使用场景

方差齐性
样本1方差 12、样本2方差 10:F = 12/10 = 1.2,df=(n1−1, n2−1)。越接近 1 越说明两总体方差齐。
ANOVA F
组间 MSB=30、组内 MSW=15:F=2.0,df=(2,27),查表临界约 3.35,未达显著。

❓ 常见问题(FAQ)

F 为何取大/小?
通常把大方差放分子使 F≥1,与单尾临界比较;方差齐性检验看双尾。
F 检验对正态敏感?
是,方差齐性检验对非正态较敏感,可考虑 Levene 检验更稳健。