Glossary
Significance Level (α)
The pre-specified probability threshold below which the null hypothesis is rejected. Common choices are 0.05 (5%) and 0.01 (1%). The significance level also equals the type I error rate — the probability of rejecting a true null hypothesis.
Definition
The pre-specified probability threshold below which the null hypothesis is rejected. Common choices are 0.05 (5%) and 0.01 (1%). The significance level also equals the type I error rate — the probability of rejecting a true null hypothesis.
Why It Matters
The significance level sets the boundary between "statistically significant" and "not significant" results, directly controlling the false-positive rate. Choosing a lower significance level (e.g., 0.01) reduces the risk of false discoveries but increases the risk of missing true effects. Researchers must balance these risks based on the consequences of each type of error in their specific context.
Example
A pharmaceutical company tests whether a new drug lowers blood pressure. Setting α = 0.05 means the company accepts a 5% chance of concluding the drug works when it actually does not. In a high-stakes safety study, the company might choose α = 0.01 to reduce the false-positive rate to 1%.
Related Terms
Software Notes
- SPSS: Set in Analyze > Compare Means / Regression dialogs; default α = 0.05
- R: Specified in hypothesis test functions; e.g.,
t.test(x, y, conf.level = 0.99)for α = 0.01 - Stata: Set via
set level 99for 99% confidence (α = 0.01); default is 95% (α = 0.05)