liveBook · Manning

Composite hypothesis

As explained in Timeless Algorithms, The Foundational Ideas

A hypothesis that permits multiple parameter values or probability models.

A composite hypothesis permits a range of values for a population parameter, rather than specifying a single value as a simple hypothesis does. It may be an alternative to a simple null hypothesis. section 4.1.1

The range can include departures in either direction or only in a specified direction. Define the hypotheses and test structure before examining data, then keep them fixed to avoid changing the rules in response to results. section 4.1.1

Examples show how the allowed range depends on the research question:

  • For beverage fills, the alternative to a mean of 12 ounces includes means both above and below 12. section 4.2.1
  • For a drug trial, the alternative to a 50% recovery probability includes probabilities greater than 50%. section 4.2.3

How Timeless Algorithms, The Foundational Ideas covers it

How deep it goes
Distinguishes range-valued from single-value hypotheses and illustrates directional alternatives without deriving test procedures.
Examples
Uses beverage fill means above or below 12 ounces and clinical-trial recovery probabilities above 50%.
What you can do after
Understand how a research question determines the range and direction of an alternative hypothesis.
How it connects
Connects composite hypotheses to simple hypotheses, null and alternative hypotheses, and one- or two-sided tests.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Timeless Algorithms, The Foundational Ideas. The section links point to the parts of the book it draws on.