F Distribution Table

Critical values of the F distribution, used in ANOVA, regression and tests that compare two variances. Pick α, then enter both degrees of freedom above to highlight the cell and get the exact value for any df.

How to read the F table

An F value has two sets of degrees of freedom, and the order matters.

  • df₁ (numerator) runs across the top. In one-way ANOVA it is the number of groups − 1.
  • df₂ (denominator) runs down the side. In one-way ANOVA it is the total sample size − the number of groups.

Each α has its own table, so choose α first. Then find the column for df₁ and the row for df₂. If your F statistic is larger than that value, the result is significant.

A worked example

A teacher compares test scores from four teaching methods, with six students in each group (24 in total). The ANOVA gives F = 4.12.

  • df₁ = 4 − 1 = 3 and df₂ = 24 − 4 = 20.
  • In the α = 0.05 table, column 3, row 20 gives 3.10. Show this in the table.
  • 4.12 is larger, so at least one method’s mean differs from the others.
  • Switch to α = 0.01: the value is 4.94, which 4.12 does not reach. So 0.01 < p < 0.05. The exact p is 0.020.

A significant F tells you the groups are not all equal. It doesn’t say which ones differ. For that, run a post-hoc test such as Tukey’s HSD, which the ANOVA calculator includes.

When your df is not in the table

Use the nearest smaller df, as with the t table. With df₂ = 50, read the df₂ = 40 row (2.84 for df₁ = 3), which is a little stricter than the exact value of 2.79. Or type 50 into the lookup box and read the exact value.

Lower-tail values

The table only lists upper-tail values. For a lower-tail value, swap the degrees of freedom and take the reciprocal:

Flower(df₁, df₂) = 1 / Fupper(df₂, df₁)

For example, the lower 5% point of F(3, 20) is 1 / 8.66 = 0.115. You need this for a two-sided test that compares two variances.

Common mistakes

  • Swapping df₁ and df₂. F(3, 20) at 0.05 is 3.10, but F(20, 3) is 8.66. Numerator df always goes across the top.
  • Halving α for ANOVA. The ANOVA F test is one-tailed by design. Use α as it is.
  • Using the total sample size as df₂. Subtract the number of groups first.

Check yourself

A one-way ANOVA has 3 groups of 10. What is the critical F at α = 0.05?

df₁ = 2 and df₂ = 30 − 3 = 27. Column 2, row 27: 3.35.

A regression with 2 predictors and 33 observations reports F = 6.2. Significant at 0.01?

df₁ = 2 (number of predictors) and df₂ = 33 − 2 − 1 = 30. The α = 0.01 table gives 5.39 at column 2, row 30. 6.2 is larger, so yes.