Reference Tables

Degrees of Freedom Table & Reference

Complete degrees of freedom (df) reference for every statistical test. Find df formulas for t-tests, chi-square, ANOVA, regression & more.

Swipe tables to see all values

What Are Degrees of Freedom?

Degrees of freedom (df) represent the number of independent values that can vary in a statistical calculation. Think of it as the number of values that are “free to change” after you’ve applied constraints (like knowing the mean).

Why do they matter? The shape of test distributions (t, χ², F) depends on df. Using the wrong df gives you the wrong critical value and the wrong conclusion.


Quick Reference: df for Every Test

Statistical TestDegrees of Freedom FormulaExample
One-sample t-testdf = n − 1n = 25 → df = 24
Paired t-testdf = n − 1 (n = number of pairs)15 pairs → df = 14
Two-sample t-test (pooled)df = n₁ + n₂ − 220 + 25 → df = 43
Welch’s t-testComplex formula (see below)Use software
Chi-square goodness-of-fitdf = k − 15 categories → df = 4
Chi-square independencedf = (r − 1)(c − 1)3×4 table → df = 6
One-way ANOVA (between)df₁ = k − 14 groups → df₁ = 3
One-way ANOVA (within)df₂ = N − k40 total, 4 groups → df₂ = 36
Simple linear regressiondf = n − 2n = 30 → df = 28
Multiple regressiondf = n − p − 1n = 50, 3 predictors → df = 46
F-test (regression)df₁ = p, df₂ = n − p − 13 predictors, n = 50 → df₁ = 3, df₂ = 46

Degrees of Freedom for T-Tests

One-Sample T-Test

df=n1df = n - 1

Example: You survey 30 students about study hours. df = 30 − 1 = 29. Look up df = 29 in the t-table.

Paired T-Test

df=n1 (where n = number of pairs)df = n - 1 \text{ (where n = number of pairs)}

Example: You measure 20 patients before and after treatment. You have 20 pairs, so df = 20 − 1 = 19.

Two-Sample T-Test (Pooled / Equal Variances)

df=n1+n22df = n_1 + n_2 - 2

Example: Group A has 15 participants, Group B has 18. df = 15 + 18 − 2 = 31.

Welch’s T-Test (Unequal Variances)

df=(s12n1+s22n2)2(s12/n1)2n11+(s22/n2)2n21df = \frac{\left(\frac{s_1^2}{n_1} + \frac{s_2^2}{n_2}\right)^2}{\frac{(s_1^2/n_1)^2}{n_1-1} + \frac{(s_2^2/n_2)^2}{n_2-1}}

This is complex — use our T-Test Calculator to compute it automatically.


Degrees of Freedom for Chi-Square Tests

Goodness-of-Fit Test

df=k1df = k - 1

where k = number of categories.

Categoriesdf
2 (e.g., yes/no)1
32
43
54
65

Example: Testing if a die is fair (6 categories). df = 6 − 1 = 5. Look up in the chi-square table.

Test of Independence / Homogeneity

df=(r1)(c1)df = (r - 1)(c - 1)

where r = rows, c = columns.

Table Sizedf
2×21
2×32
2×43
3×34
3×46
4×49
4×512
5×516

Example: A 3×4 contingency table. df = (3 − 1)(4 − 1) = 2 × 3 = 6.


Degrees of Freedom for ANOVA

One-Way ANOVA

Sourcedf FormulaExample (4 groups, N = 40)
Between groupsdf₁ = k − 14 − 1 = 3
Within groupsdf₂ = N − k40 − 4 = 36
Totaldf_total = N − 140 − 1 = 39

Check: df_between + df_within = df_total → 3 + 36 = 39 ✓

Look up F-critical with df₁ and df₂ in the F-table.

Two-Way ANOVA

Sourcedf
Factor Aa − 1
Factor Bb − 1
Interaction (A×B)(a − 1)(b − 1)
Within (Error)N − ab
TotalN − 1

Degrees of Freedom for Regression

Simple Linear Regression (1 predictor)

Sourcedf
Regression1
Residualn − 2
Totaln − 1

Multiple Regression (p predictors)

Sourcedf
Regressionp
Residualn − p − 1
Totaln − 1

Example: Predicting exam score from 3 variables (study hours, sleep, attendance) with n = 50 students.

  • Regression df = 3
  • Residual df = 50 − 3 − 1 = 46
  • F-test: use df₁ = 3, df₂ = 46 in the F-table

Worked Example: Choosing the Right df

Problem: You want to test whether male and female students prefer different types of music (Rock, Pop, Classical, Jazz). You survey 200 students.

  1. Test type: Chi-square test of independence (two categorical variables)
  2. Table size: 2 rows (gender) × 4 columns (music type)
  3. df = (2 − 1)(4 − 1) = 3
  4. Critical value at α = 0.05: Look up df = 3 in the chi-square table7.815
  5. If your calculated χ² > 7.815, gender and music preference are significantly associated.

Common Mistakes with Degrees of Freedom

  1. Using n instead of n − 1 for one-sample t-tests
  2. Confusing one-way and two-way ANOVA df — two-way requires interaction terms
  3. Forgetting to subtract 1 in chi-square goodness-of-fit
  4. Using the wrong df for Welch’s t-test — it’s not simply n₁ + n₂ − 2
  5. Counting total N instead of pairs for paired t-tests

Frequently Asked Questions

Why do we subtract 1 from the sample size?

When calculating from a sample, knowing the mean constrains one value. If you know the mean of 5 numbers and 4 of them, the 5th is determined. So only n − 1 values are “free” — hence df = n − 1.

What happens as degrees of freedom increase?

The t-distribution approaches the normal (z) distribution. With df > 30, t-values are very close to z-values. At df = ∞, they are identical.

Can degrees of freedom be a decimal?

Yes! Welch’s t-test often gives non-integer df (e.g., 23.7). Statistical software uses the exact value; for table lookups, round down to be conservative.

What if my df isn’t in the table?

Use the next smaller df value listed in the table for a conservative result, or use statistical software for the exact p-value. Our interactive tables accept any df value.


Statistical Tables
Z-table, t-table, chi-square & F-table — free printable reference tables.
View Tables →

Need to calculate values?

Use our free calculators to compute statistics without looking up tables manually.

Browse Calculators →