What Are Degrees of Freedom? Definition
Degrees of freedom (df) indicate the number of independent pieces of information used to calculate a statistic. In other words, they represent the maximum number of independent values that are free to vary in a dataset. This is generally calculated by subtracting one from the sample size. Degrees of freedom are critical for validating statistical tests such as chi-square tests, ANOVA tests, t-tests, and F-tests.
The number of degrees of freedom varies with the sample size:
- If the sample size (n) is small, degrees of freedom will also be small.
- If the sample size (n) is large, degrees of freedom will also be large.
💡 Degrees of freedom are connected to sample size, but they are not the same thing. The degrees of freedom are always fewer than the sample size.
How to Find Degrees of Freedom – Formulas
The formula depends on the type of statistical test you are performing:
1-Sample t-test
df = N − 1
where N is the total number of subjects/values in the sample.
2-Sample t-test (Equal Variances)
df = N₁ + N₂ − 2
where N₁ and N₂ are the sizes of the two independent samples.
Welch's t-test (Unequal Variances)
An approximation of degrees of freedom is computed via the Welch–Satterthwaite equation:
df ≈ (s₁²/N₁ + s₂²/N₂)² / [(s₁²/N₁)²/(N₁−1) + (s₂²/N₂)²/(N₂−1)]
where s₁² and s₂² are the sample variances of the two groups.
ANOVA
- dfbetween = k − 1 (where k = number of groups)
- dfwithin = N − k (where N = total observations)
- dftotal = N − 1
Chi-Square Test of Independence
df = (rows − 1) × (columns − 1)
Applied to a contingency table with a given number of rows and columns.
Example: How to Calculate Degrees of Freedom by Hand
Suppose you are performing a two-sample t-test with equal variances comparing two groups of students: Group A (n₁ = 25) and Group B (n₂ = 30).
- Identify the formula: df = N₁ + N₂ − 2
- Substitute: df = 25 + 30 − 2 = 53
You would then look up a t-distribution table with df = 53 to find the critical t-value at your chosen significance level (e.g., α = 0.05).
Frequently Asked Questions (FAQs)
What does a higher degrees of freedom mean?
A higher df generally means your test statistic follows a distribution closer to the standard normal (Z) distribution, resulting in more statistical power and narrower confidence intervals.
Can degrees of freedom be a decimal?
Yes — in Welch's t-test the df is an approximation and is typically a non-integer. You can either use the exact decimal value with software or round down to the nearest integer to be conservative when consulting printed t-tables.
Why do we subtract 1 from the sample size?
When estimating the population mean from a sample, one constraint (the sample mean) is imposed on the data. This reduces the number of values that are truly "free to vary" by one, giving df = n − 1.
How are degrees of freedom used in practice?
After computing your test statistic (t, F, χ²), you use the degrees of freedom to look up the corresponding critical value or p-value in a statistical table or software. This determines whether your result is statistically significant at a chosen confidence level.