Properties and Applications of Chi-Squared
You'll learn how the chi-squared distribution's mathematical properties make it indispensable for hypothesis testing and interval estimation.
Core Properties
The chi-squared distribution with k degrees of freedom, denoted χ²(k), arises when you sum k independent squared standard normal variables. This simple origin yields powerful properties:
- Non-negativity: Since it's a sum of squares, χ² values are always ≥ 0
- Right-skewed shape: The distribution is asymmetric, skewing right, especially for small k
- Mean and variance: E[χ²(k)] = k and Var(χ²(k)) = 2k
- Additivity: If X ~ χ²(k₁) and Y ~ χ²(k₂) are independent, then X + Y ~ χ²(k₁ + k₂)
- Convergence: As k increases, χ²(k) approaches normality (by the Central Limit Theorem)
Applications in Statistical Inference
Hypothesis Testing
The chi-squared distribution is the backbone of several critical tests:
- Chi-squared goodness-of-fit test: Tests whether observed categorical data follows an expected distribution
- Chi-squared test of independence: Determines if two categorical variables are independent in contingency tables
- Variance testing: Tests hypotheses about population variance, since (n-1)s²/σ² ~ χ²(n-1)
Confidence Intervals
For sample variance s² from a normal population, we use chi-squared critical values to construct confidence intervals for the population variance σ². The interval bounds come from the percentiles of the χ²(n-1) distribution, accounting for its asymmetry (unlike symmetric normal intervals).
Key Takeaway: The chi-squared distribution's properties—especially its connection to squared normals and additivity—make it the natural choice for analyzing variance and testing categorical relationships in statistics.