Course contentsShow
← Data Science
Lesson 310 of 2,1459. Hypothesis Testing FoundationsFree lesson

Writing Hypotheses for Different Parameters

Practice formulating hypotheses for means, proportions, differences, and other population parameters.

Writing Hypotheses for Different Parameters

You'll learn: How to correctly write null and alternative hypotheses for different types of population parameters like means, proportions, and differences.

The Core Idea

Just as you wouldn't use the same key for every lock, different research questions require different hypothesis statements. The parameter you're testing (mean, proportion, difference, etc.) determines how you write your hypotheses—but the logic stays the same: H₀ always claims "no effect" or "no difference," while H₁ claims something is happening.

Common Parameter Types and Their Hypotheses

Testing a Single Mean (μ)

  • H₀: μ = some specific value
  • H₁: μ ≠ some value (two-sided) or μ > value or μ < value (one-sided)
  • Example: Is the average customer wait time 5 minutes?

Testing a Single Proportion (p)

  • H₀: p = some specific value
  • H₁: p ≠ some value (or p > value, or p < value)
  • Example: Is the conversion rate 20%?

Testing Difference Between Two Means (μ₁ - μ₂)

  • H₀: μ₁ - μ₂ = 0 (or equivalently, μ₁ = μ₂)
  • H₁: μ₁ - μ₂ ≠ 0 (or μ₁ > μ₂, or μ₁ < μ₂)
  • Example: Do two teaching methods produce different test scores?

Testing Difference Between Two Proportions (p₁ - p₂)

  • H₀: p₁ - p₂ = 0 (or p₁ = p₂)
  • H₁: p₁ - p₂ ≠ 0 (or p₁ > p₂, or p₁ < p₂)
  • Example: Do two ads have different click-through rates?

Testing Paired Differences (μ_d)

  • H₀: μ_d = 0 (mean difference is zero)
  • H₁: μ_d ≠ 0 (or μ_d > 0, or μ_d < 0)
  • Example: Does a training program improve employee performance?

Key Takeaway: The structure of your hypotheses mirrors your research question—identify the parameter (mean, proportion, difference), then state H₀ as "no effect/no difference" and H₁ as the claim you're investigating.