Student’s t Distribution — $ t_\nu $

The Student’s t distribution arises as a ratio of a Normal variable and the square root of an independent Chi-square variable. It is the fundamental distribution for mean inference with unknown variance and a canonical example of a pivot.


Definition

Degrees of freedom:

  • $ \nu \in \mathbb{N} $

Support:
\(x \in \mathbb{R}\)

PDF:
\(f(x) = \frac{\Gamma\!\left(\frac{\nu+1}{2}\right)} {\sqrt{\nu\pi}\,\Gamma\!\left(\frac{\nu}{2}\right)} \left(1+\frac{x^2}{\nu}\right)^{-(\nu+1)/2}\)


Construction from Normal and Chi-square

Let \(Z \sim \mathcal{N}(0,1), \quad U \sim \chi^2_\nu, \quad Z \perp U.\)

Then \(T = \frac{Z}{\sqrt{U/\nu}} \sim t_\nu\)

This representation is the definition used in proofs.


Relationship to Other Distributions

Connection to Chi-square

  • Denominator involves $ \chi^2_\nu $
  • Degrees of freedom come from variance estimation

Connection to Normal

\[t_\nu \Rightarrow \mathcal{N}(0,1) \quad \text{as } \nu \to \infty\]

Heavier tails for small $ \nu $.


Moments

  • Mean: $ 0 $ for $ \nu > 1 $
  • Variance: \(\operatorname{Var}(T) = \frac{\nu}{\nu-2}, \quad \nu>2\)

Higher moments exist only for sufficiently large $ \nu $.


Symmetry

\[T \overset{d}{=} -T\]

All odd moments (when defined) are zero.


Pivotal Quantity (Key Use)

If \(X_1,\dots,X_n \stackrel{iid}{\sim} \mathcal{N}(\mu,\sigma^2),\) then \(\frac{\bar{X}-\mu}{S/\sqrt{n}} \sim t_{n-1}\)

This is a pivot, free of unknown parameters.


Role in Inference

Used for:

  • Confidence intervals for $ \mu $
  • Hypothesis testing when variance is unknown
  • Regression coefficient inference

Likelihood Perspective

The t distribution arises from integrating out the variance in a Normal–Inverse-Gamma model.

Interpretation:

  • t is a scale mixture of Normals
  • Explains heavy tails

Tail Behavior

\[\mathbb{P}(|T|>x) \sim C x^{-\nu} \quad \text{as } x \to \infty\]

Heavier tails than Normal.


Multivariate Extension (Recognition Only)

  • Multivariate t arises from Normal with random scale
  • Used in robust modeling

(No formulas required for prelims.)


Key Theorems and Facts (Prelim-Relevant)

  • Ratio of Normal and Chi-square
  • Pivot for mean with unknown variance
  • Converges to Normal
  • Heavy tails
  • Scale-mixture interpretation

Exam Takeaways

  • Unknown variance ⇒ t
  • Degrees of freedom = sample size − 1
  • Think “Normal divided by Chi-square”
  • Heavier tails protect against variance uncertainty
  • Central object for classical inference

Comments