Back to questions

R Squared Misuse [Amazon Statistics Interview Question]


Suppose the Product Manager on the Amazon Marketplace Fraud Team asks you to analyze the relationship between the chance a seller is fraudulent seller, and some other variable about the seller.

They took Statistics 101 a decade ago, so they confidently tell you "Don't just describe the relationship qualitatively - I'm a numbers person! Be sure to include the R-Squared (R2R^{2}) in your report!"

What are some ways R2R^2 could be misused, or be misleading, or not tell the PM the whole picture?