← Meta Interview Insights

Meta·Data Scientist·Technical Phone Screen·Intermediate

Intermediate
May 2026

Summary

Meta data scientist interview with a stats-heavy question on distributions and the central limit theorem. Pretty theoretical, felt more like a grad school oral exam than a product interview.

Questions Asked (1)

Q1

Sketch a right-skewed distribution of users' daily comment counts and mark where the mean, median, and 95th percentile fall. Then suppose you repeatedly sample random groups of users and compute each group's average comment count. What does that new distribution look like, and what happens to the mean, median, and 95th percentile?

Product Analytics & MetricsA/B Testing & Experimentation
Author's notes

The sketch part was fine, mean to the right of median, long right tail, 95th percentile way out there.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

First, sketch a right-skewed distribution and correctly place the mean, median, and 95th percentile. Then explain that the sampling distribution of the sample mean will be approximately normal (by the Central Limit Theorem) with the same mean but smaller variance, and discuss how the median and 95th percentile behave differently.

Pro tip: Emphasize that the mean of the sampling distribution equals the population mean, but the median and 95th percentile of the sampling distribution are not equal to their population counterparts; they converge to the population mean as sample size increases. This shows deep understanding of the CLT and its implications.

1. Sketch the population distribution

Draw a right-skewed distribution of daily comment counts. Mark the mean to the right of the median, and the 95th percentile far to the right, indicating a long tail.

2. Explain the sampling distribution

Describe that when repeatedly sampling groups of users and computing the average, the distribution of these averages will be approximately normal (by the CLT) with the same mean as the population but reduced variance (standard error = σ/√n).

3. Compare mean, median, and 95th percentile

State that the mean of the sampling distribution equals the population mean. The median and 95th percentile of the sampling distribution will be closer to the mean than in the population, and as sample size increases, they converge to the population mean.

4. Discuss implications for product analytics

Relate this to A/B testing: when comparing average comment counts between groups, the sampling distribution of the mean is normal, enabling standard hypothesis tests. However, metrics like the 95th percentile require different methods (e.g., bootstrapping) due to non-normality.

Key Points to Mention

  • Right-skewed distribution: mean > median, long right tail.
  • 95th percentile is far to the right, representing heavy users.
  • Central Limit Theorem: sampling distribution of the mean is approximately normal for large n.
  • Mean of sampling distribution equals population mean; standard error decreases with sample size.
  • Median and 95th percentile of sampling distribution are not equal to population values; they converge to the mean as n increases.
  • Implications for A/B testing: use t-tests for means, but consider bootstrapping or quantile regression for percentiles.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.