Loading
Interactive hypothesis testing visualization showing sampling distribution, rejection regions, critical values, test statistic, and decision logic. Understand how significance level α defines rejection regions and how to interpret test results.
Distribution
Test Type
Hypothesis testing uses a test statistic to determine whether to reject the null hypothesis (H₀). The test statistic measures how many standard errors the sample mean is from the hypothesized population mean. For a Z-test: z = (x̄ - μ₀) / (σ / √n). For a T-test: t = (x̄ - μ₀) / (s / √n). The distribution of this test statistic under H₀ is either standard normal (Z) or t-distributed.
The sliders, formulas and analytics view need more room than a phone screen can give them. Open this chart on a desktop or larger tablet to use the full interactive experience.
Avoid these frequent errors
Confusing "fail to reject H₀" with "accept H₀" (we never accept, only fail to reject)
Using wrong tails: two-tailed for ≠, right-tailed for >, left-tailed for <
Forgetting to divide α by 2 for two-tailed tests when using tables
Using z-test when σ is unknown and n is small (should use t-test)
Confusing significance level α with p-value (α is predetermined, p-value is calculated)
Thinking statistical significance means practical importance
Not adjusting for multiple testing (increases Type I error)
Confusing one-sided and two-sided critical values
Strategic insights for success
MEMORIZE: Z-critical values: ±1.645 (10%), ±1.96 (5%), ±2.58 (1%) for two-tailed
MEMORIZE: For one-tailed at 5%: Z-critical = 1.645 (or -1.645 for left tail)
Test statistic in rejection region → Reject H₀
p-value < α → Reject H₀ (both methods give same answer)
Small p-value = strong evidence against H₀
CFA loves: "Is this result statistically significant at 5% level?"
Watch for: z-test (σ known) vs t-test (σ unknown, or small sample)
Common setup: Given sample mean, hypothesized mean, std dev, sample size → calculate test statistic
Decision wording: "Reject H₀" or "Fail to reject H₀" (NEVER "accept")
Quick check: If |test statistic| > |critical value|, reject H₀ (for two-tailed)