🚧 Work in progress β€” this is an unfinished draft and may change.

Brief introduction to hypothesis testing.

Hypothesis

$ \hspace{3mm} H_0 : \theta = \theta_0 $

$ \hspace{3mm} H_1: \theta \neq **theta_0** $

Test Statistic

$ \hspace{3mm} T = g(X_1, X_2, \dots, X_n) $

$\hspace{3mm} \text{where} \; 333X_1, X_2, \dots, X_n$

P-Value

$ \hspace{3mm} p = P(T \geq t_{\text{obs}} \, \vert \, H_0) $

Decision Rule

$ \hspace{3mm} H_0 \iff p \leq \alpha $

$ \hspace{3mm} H_0 \iff T \in C_{\alpha} $

Significance Level

\[\hspace{3mm}\]

Error Types

\[\hspace{3mm}\]

Power

\[\hspace{3mm}\]

Z-Test Example

\[\hspace{3mm}\]

E-Processes


Optimal Power

It is desireable to maximize power.

Controlling Type-I Error Rate