Importance of central limit theorem

The Central Limit Theorem (CLT) states: If $X_1, X_2, \dots, X_n$ are i.i.d. random variables from a population with mean $\mu$ and finite variance $\sigma^2$, then the sample mean $\bar{X} = \frac{1}{n} \sum_{i=1}^n X_i$ has a distribution that approaches the normal distribution as $n \to \infty$:

Unlike the population itself, which may be skewed, uniform, or discrete (e.g., Bernoulli), the sampling distribution of $\bar{X}$ is approximately normal for large $n$ (typically $n \geq 30$), regardless of the original shape.
This enables normal-based methods even for non-normal data.

CLT justifies using the normal distribution (or $z$-/$t$-distributions) for:

• Confidence intervals: $\bar{x} \pm z_{\alpha/2} \frac{s}{\sqrt{n}}$.
• Hypothesis tests for means: $Z = \frac{\bar{x} - \mu_0}{s/\sqrt{n}}$.
  Without CLT, inference would require exact distributions, which are often unknown or complex.

Many real-world phenomena (heights, errors, totals) are sums/averages of many small effects, so they follow normals due to CLT. This is why the bell curve appears "everywhere."

• Polling: Survey averages approximate normals for election predictions.
• Quality Control: $\bar{X}$-charts monitor process means.
• Finance: Portfolio returns (sums of assets) are near-normal.
• Machine Learning: Bootstrap and resampling rely on CLT-like properties.

Requires finite variance and i.i.d.; fails for heavy-tailed distributions (e.g., Cauchy). Use $n > 30$ rule cautiously.

The CLT is the cornerstone of modern statistics, bridging theory to practice.