Test whether two user populations differ
Company: Amazon
Role: Applied Scientist
Category: Machine Learning
Difficulty: medium
Interview Round: Onsite
## Problem
You are given two groups of users:
- Group A: North America users
- Group B: Europe users
Each user has a vector of **continuous** features (e.g., session duration, click-through rate, purchase conversion rate, etc.).
### Task
Describe how you would determine whether the two groups differ **significantly in their overall multivariate distribution**.
Your answer should cover:
- Hypotheses (what does “same distribution” mean?)
- Preprocessing/normalization considerations
- One or more statistical tests or modeling approaches
- How you would report significance and effect size
- Pitfalls (multiple testing, confounding, non-IID, heavy tails, missing data)
Overview: This question evaluates competency in statistical inference and multivariate analysis, covering hypothesis formulation about equal distributions, preprocessing and normalization considerations, choice of statistical tests or modeling approaches, and interpretation of significance and effect size.
Read the full Amazon Applied Scientist interview experience this question came from
