Every Statistic Hypothesis involves two opposite statements about a population parameter Θ (like 𝜇,𝜋,𝜎).

Concepts

Null hypothesis (Ho)​:

The “status quo” or a statement of no effect / no difference.

  • Usually contains an equality: ,,=.
  • We assume it is true unless sample evidence strongly contradicts it.

Alternative hypothesis, denoted (or Ha​):

The opposite of 𝐻0, what we suspect or want to prove.

  • Contains a strict inequality: >,<,.
  • Key rule: 𝐻0​ is always stated first, then 𝐻𝑎​.

Examples from the text:

  1. Suspect that the illiteracy rate in Peru is at most 5%:
    1. 𝐻0:𝜋0.05
    2. 𝐻𝑎:𝜋>0.05
  2. Suspect the average exam score is below 13:
    1. 𝐻0:𝜇13
    2. 𝐻𝑎:𝜇<13

How to Formulate

Often it is easier to identify 𝐻𝑎 first (what you want to prove), then 𝐻0 is simply the opposite.

Two common scenarios

H_a as the research hypothesis

When you want to demonstrate that something is better, more effective, higher, etc.

  • 𝐻𝑎 states the improvement; 𝐻0 states “no improvement” or “not better.”

Example: A new teaching method is believed to be better than the current one.

  • 𝐻0: The new method is not better.
  • 𝐻𝑎: The new method is better.

Ho as an assumption to be challenged

Often used in quality control or label verification.

  • 𝐻0 contains the claimed or required value (with =, , or ).
  • 𝐻𝑎 is the opposite (e.g., under‑filling, over‑filling, not meeting specification).\

Example: A soda bottle label says 0.5 L

  • 𝐻0:𝜇=0.5 (the claim is correct)
  • 𝐻𝑎:𝜇0.5 (the claim is false)