What is Levene's Test?

Levene's test is an inferential statistic used to assess the equality of variances for a variable calculated for two or more groups. This property is known as "Homogeneity of Variance" (Homoscedasticity). The test calculates the absolute deviations of data points from the group center (Mean or Median) and performs a one-way ANOVA on those deviations.

When to use it?

Statistical tests like the Independent T-Test and ANOVA assume that the variance across populations is equal. You use Levene's test to:

• Validate ANOVA Assumptions: If Levene's is significant (p < 0.05), you cannot assume equal variances.
• Choose the Right Correction: If variances are unequal, you must use corrections like Welch’s t-test or Welch’s ANOVA.

How our tool works (Accuracy & Stack)

This tool utilizes Python's `scipy.stats.levene` on Google Cloud, offering precision matching R and SAS. We offer flexibility by allowing you to choose the center:

• Mean: For symmetric, normal distributions.
• Median: For skewed distributions (This technically makes it the Brown-Forsythe method, though often still accessed via Levene's logic).

Uniquely, our tool visualizes the variance itself:

• Absolute Deviations Plot: A specialized chart showing the spread of deviations from the center, making it easy to see which group has higher variability.
• Box & Violin Plots: To visualize the raw distribution spread.

Results and charts are available in the downloadable PDF Report.

Input Requirements & Limitations

To maintain high-speed performance for calculations and plotting:

• Row Limit: Supports up to 5,000 rows per group.
• Column Limit: Supports up to 10 groups (columns).
• Total Data Points: Maximum 50,000 data points.

Import your dataset via CSV for instant analysis. No data is stored.

Comparison: Levene's vs. Bartlett's Test

Bartlett's test is another method for testing equal variances, but it is highly sensitive to departures from normality. If your data is not perfectly normal, Bartlett's may give a "false positive." Levene's test is far more robust to non-normal data, making it the safer choice for most real-world datasets.