How to Compare Groups in Your Data
Learn how to compare groups (regions, departments, products) in your data with statistical rigor. Understand when differences are real vs random noise.
Step-by-step guide
Identify your groups
Groups can be regions, departments, products, time periods, treatment arms, or any categorical column in your data.
Upload your data
Upload your CSV with the group column and outcome columns. OakPrism auto-detects group variables.
Review comparisons
OakPrism runs appropriate statistical tests and reports which differences are significant with effect sizes.
Understand the nuance
OakPrism checks for Simpson's paradox and confounders to ensure subgroup trends don't mislead.
Why Group Comparisons Matter
Most business insights come from comparisons: which product sells better, which region is underperforming, which treatment works. But not all differences are meaningful — some are just random noise.
Statistical Significance Explained
A statistically significant difference means the gap you see is unlikely to be caused by chance alone. OakPrism automatically runs the appropriate test (t-test, ANOVA, chi-square, Mann-Whitney) based on your data type.
Practical Significance
Even statistically significant differences might be too small to matter. OakPrism reports both statistical significance AND effect size so you know which differences are worth acting on.