Statistics Interview Prep

• Explain fundamental statistics concepts conversationally
• Walk a reviewer through a method choice end to end
• State the limits of p-values and significance without prompting

• Analysts interpret tests and set up experiments — statistical literacy is the job.
• The p-value is the most over-trusted number in the industry; showing you know why is a strong signal.
• Method defensibility turns 'I got a p<0.05' into a decision maker trusts.

The number is only useful with its context: which wrench, which spec. A p-value only matters when you know the test, the assumption, and what misuse would look like.

• Master the core concepts as conversational explanations, not definitions.
• Practice choosing a method from a story: what is the outcome, what is the comparison?
• Learn to walk a hypothesis test end to end — setup, power, result, caveats.
• Drill the common probes: 'is significance enough?', 'what about the base rate?'

• An interviewer gives a funnel drop-off and asks what test and interpretation you would run.
• A product case asks whether a 1% lift with p=0.04 is a green light — the answer is not automatic.
• Sample-size reasoning is probed with: 'how many users to detect a 5% lift?'

• Answer in decision language: what would you do differently with this result?
• Volunteer the caveats before they are asked — it signals depth.
• Practice 'What would make my conclusion wrong?' on every past project.

• Confusing statistical significance with practical importance.
• Forgetting the base rate when interpreting a probability.
• Citing formulas without the logic of why the method fits the question.

• If you freeze on a concept, narrate the intuition until it unlocks — interviewers reward recovery.
• If you catch an error in your own reasoning, say so and correct it; correction is a feature, not a flaw.