Statistical tables

F Distribution (Snedecor)

Critical values F(α, ν₁, ν₂) for the most commonly used significance levels. Rows are the denominator degrees of freedom (ν₂) and columns the numerator degrees of freedom (ν₁). Click a cell to see the value.

Table for α = 0.10: each cell shows F such that P(F > f) = 0.10. Rows: ν₂ (denominator). Columns: ν₁ (numerator). The symbol ∞ represents a very large degrees of freedom (≈ 10,000).
Click a cell to see the critical value.

F(α=0.10, ν₁, ν₂) — upper 10% tail

Table for α = 0.05: each cell shows F such that P(F > f) = 0.05. This is the most commonly used table in ANOVA and equality-of-variances tests.
Click a cell to see the critical value.

F(α=0.05, ν₁, ν₂) — upper 5% tail

Table for α = 0.025: each cell shows F such that P(F > f) = 0.025. Used in two-sided tests at the 5% level on variance ratios.
Click a cell to see the critical value.

F(α=0.025, ν₁, ν₂) — upper 2.5% tail

Table for α = 0.01: each cell shows F such that P(F > f) = 0.01. Used in highly significant tests.
Click a cell to see the critical value.

F(α=0.01, ν₁, ν₂) — upper 1% tail

How to read the table: each cell shows P(F ≤ f) for the selected numerator (ν₁) and denominator (ν₂) degrees of freedom. The row gives the integer part, the column the first decimal.
Click a cell to see the value.

P(F ≤ f, ν₁, ν₂) for f from 0.0 to 15.9

How to use these tables

F Distribution (Snedecor)

If X₁ ~ χ²(ν₁) and X₂ ~ χ²(ν₂) are independent, then F = (X₁/ν₁) / (X₂/ν₂) follows an F distribution with ν₁ and ν₂ degrees of freedom. Its density function is:

\( f(f;\nu_1,\nu_2) = \dfrac{\sqrt{\dfrac{(\nu_1 f)^{\nu_1}\,\nu_2^{\nu_2}}{(\nu_1 f+\nu_2)^{\nu_1+\nu_2}}}}{f\,B\!\left(\tfrac{\nu_1}{2},\tfrac{\nu_2}{2}\right)}, \quad f > 0 \)

The mean is ν₂/(ν₂ − 2) for ν₂ > 2. It has positive skew.

Critical values

The cell in row ν₂ and column ν₁ shows the quantile F such that P(F > f) = α. For an ANOVA at the 5% level with ν₁ = 3 groups and ν₂ = 20 residuals, look up the α = 0.05 table → F = 3.098.

  • For equality-of-variances tests (two-sided), use α/2 in the table and reject if F > F(α/2) or F < 1/F(α/2).
  • As ν₁ → ∞ and ν₂ → ∞, the F distribution tends to 1.

Related calculators