Statistics & Math Interview Questions

Practice 592 real statistics and probability interview questions from Meta, Capital One, Google, Optiver and Uber. They cover conditional probability and Bayes, expected value, which distribution applies and why, hypothesis testing and what a p-value does not tell you, confidence intervals, the central limit theorem, sampling bias, correlation against causation, and the brainteasers trading desks still open with. 471 come from Data Scientist loops. The quant firms account for the sharper end, where questions are timed, asked out loud, and followed up until you reach the edge of what you actually know. 269 were asked in technical screens and 168 onsite, with 51 more set as take-home work. Each question records the company, role and round it came from, with a worked solution that shows the setup rather than only the final number.

592 Questions 99 Companies09.17.2026
Showing 12 results

Frequently Asked Questions

How difficult are Statistics & Math interview questions?
Difficulty varies by role and seniority, from straightforward probability and descriptive stats at entry level to rigorous experiment design, inference edge cases, and modelling tradeoffs at senior levels. Interviewers evaluate statistical thinking more than raw algebra: clear assumptions, correct use of distributions, interpretation of effect sizes, and risk of bias matter most. For data roles at tech companies, a strong candidate demonstrates both calculation fluency and applied judgment, turning numeric answers into product or business implications. Expect interviewers to probe edge cases, sample assumptions, and the consequences of violated model conditions.
Where do Statistics & Math questions appear in a typical interview loop and which companies weight them heavily?
Statistics and math usually appear in technical phone screens and in one or two onsite analytics or modeling rounds. Early screens or take-home assessments test core probability, hypothesis testing, and quick inference; onsite rounds dig into experiment design, causal reasoning, and metric construction. Companies with heavy emphasis include Meta, Capital One, and Google, with Amazon and Uber also commonly testing statistical judgment depending on the team. These rounds are often 45 to 60 minutes, and they are commonly paired with SQL or product-analytics questions so candidates must connect math to real metrics and decisions.
How long should I prepare for Statistics & Math interviews?
Preparation time depends on background and urgency. Candidates with recent statistics experience typically need two to six weeks of focused review to sharpen experimental design and applied inference. People refreshing basics can prepare in one to two weeks of intensive practice. Those switching from nonstatistical roles should budget six to twelve weeks to rebuild foundations, practice problems, and translate results into product insights. Effective prep mixes problem practice, targeted review of weak subtopics, and mock interviews that force you to explain assumptions, practical significance, and limitations under time pressure.
What key subtopics in Statistics & Math should I study for these interviews?
Prioritize probability and distributions, central limit theorem, hypothesis testing and confidence intervals, statistical power and sample size, and multiple-testing corrections. Study regression and generalized linear models, bias versus variance, basic causal inference and A/B testing design, uplift and metric sensitivity, and nonparametric approaches such as bootstrapping. Also review likelihood intuition, Bayesian versus frequentist framing, handling missing data and confounders, and practical diagnostics like residual analysis and robustness checks. Employers look for both the math and the ability to translate results into product or business recommendations.
What standout tips and common pitfalls should candidates know for Statistics & Math interviews?
Always state assumptions before computing, and report practical effect sizes alongside p values. Sanity check answers with orders of magnitude and units, and discuss power and sample size when relevant. Avoid common traps: confusing correlation with causation, ignoring multiple comparisons, and relying solely on p values without context. Practice explaining results to nontechnical interviewers, and tie statistical choices to product impact. Finally, show tradeoff thinking: when a simple estimator is preferable to a complex model, and how you would monitor and validate your metric in production to catch bias or drift.

Explore more Statistics & Math interview questions

Jump straight to Statistics & Math questions at a specific company or for a specific role.

By company
Meta89Capital One57Google33Amazon21Optiver21Uber18Citadel18Sig16
By role
Data Scientist495Software Engineer48Quantitative Researcher29