Course contentsShow
← Data Science
Lesson 323 of 2,1459. Hypothesis Testing FoundationsFree lesson

What is a Significance Level (α)?

Understanding α as the threshold probability for rejecting the null hypothesis and its role in decision-making.

What is a Significance Level (α)?

What you'll learn: You'll understand how the significance level (α) acts as your decision threshold in hypothesis testing, determining when evidence is strong enough to reject the null hypothesis.


The Decision Line in Hypothesis Testing

The significance level, denoted by the Greek letter α (alpha), is a predetermined probability threshold you set before conducting a hypothesis test. It defines how much "unusual" evidence you need to reject the null hypothesis (H₀).

Think of α as the prosecutor's burden of proof. Just as "beyond reasonable doubt" sets a standard for conviction, α sets your standard for rejecting H₀. If your p-value falls below α, the evidence is strong enough to reject the null hypothesis.

Common Significance Levels

The most commonly used significance levels are:

  • α = 0.05 (5%): The standard in most scientific research
  • α = 0.01 (1%): For situations requiring stronger evidence
  • α = 0.10 (10%): For exploratory analyses where you're more tolerant of risk

How α Works in Practice

When you calculate a p-value from your test statistic, you compare it to α:

  • If p-value < α: Reject H₀ (the result is "statistically significant")
  • If p-value ≥ α: Fail to reject H₀ (insufficient evidence)

For example, with α = 0.05 and a calculated p-value of 0.03, you would reject H₀ because 0.03 < 0.05. But if your p-value were 0.08, you'd fail to reject H₀.

Why α Matters

Setting α controls your willingness to make a Type I error (falsely rejecting a true null hypothesis). A smaller α makes you more conservative, requiring stronger evidence before claiming a significant result.

Key Takeaway: The significance level (α) is your pre-determined threshold for decision-making in hypothesis testing—if your p-value is below α, you have sufficient evidence to reject the null hypothesis.