Input Data
Enter your numeric data (one value per line or comma-separated)
Q-Q Plot
Points should follow the diagonal line if data is normally distributed
Histogram with Normal Curve
Distribution should match the overlaid normal curve
Help
How to Use
- Enter your numeric data (comma or newline separated)
- Or generate sample data with different distributions
- Set your desired significance level (typically 0.05)
- Click "Run Normality Tests"
- Review the overall assessment and individual test results
- Examine Q-Q plot and histogram for visual confirmation
Understanding the Tests
- Shapiro-Wilk: Most powerful test for n < 2000; null hypothesis: data is normal
- Anderson-Darling: Gives more weight to tails; good for detecting tail deviations
- Kolmogorov-Smirnov: Tests maximum difference between empirical and theoretical CDF
- D'Agostino-Pearson: Based on skewness and kurtosis; omnibus test
Interpreting Results
- Green (Normal): Most tests accept normality (p > α)
- Yellow (Questionable): Mixed results; some tests reject, others accept
- Red (Non-Normal): Most tests reject normality (p < α)
- P-value: Probability of observing the data if it were normal; lower values indicate non-normality
Skewness and Kurtosis
- Skewness = 0: Symmetric distribution (like normal)
- Skewness > 0: Right-skewed (tail extends right)
- Skewness < 0: Left-skewed (tail extends left)
- Excess Kurtosis = 0: Normal tail behavior
- Excess Kurtosis > 0: Heavy tails (more outliers than normal)
- Excess Kurtosis < 0: Light tails (fewer outliers than normal)