##### Scenario Analyzing distribution of user comment counts and the effect of sampling. ##### Question Sketch the distribution of individual users’ daily comment counts when it is right-skewed and mark mean, median, and 95th percentile positions. After randomly sampling many user groups and computing each group’s average comments, describe the resulting distribution and how the mean, median, and 95th percentile change. ##### Hints Invoke Central Limit Theorem; sample means trend toward normal, mean stays constant, higher percentiles shrink, median approaches mean.

Evaluates statistical reasoning about right-skewed user comment counts and sampling distributions. Strong answers place median, mean, and p95 correctly, then use the Central Limit Theorem to explain why group averages become less variable and more symmetric.

How do I approach Statistics & Math interview questions?

Statistics & Math questions require understanding of core concepts and practice. PracHub provides solutions with explanations to help you master statistics & math interviews.

What difficulty level is this interview question?

This is a medium difficulty Statistics & Math question, commonly asked during Onsite rounds at Meta.

What role is this question designed for?

This question is commonly asked for Data Scientist candidates at Meta during technical interviews.

Analyze User Comment Distribution and Sampling Effects

You are analyzing daily comment counts per user. The individual user-level distribution is right-skewed: many users make zero or few comments, while a small number make many comments.

Constraints & Assumptions

Treat the outcome as a nonnegative count.
Discuss both the original per-user distribution and the sampling distribution of group averages.
Use the Central Limit Theorem carefully: it applies to averages, not to individual counts becoming normal.
Assume random groups of equal size unless stated otherwise.

Clarifying Questions to Ask Guidance

Are zero-comment users included?
What is the time window: daily, weekly, or monthly comments?
Are users sampled independently, or are there clusters such as groups or markets?
Is the goal descriptive analysis, experiment planning, or anomaly detection?

Part 1 - Sketch the User Distribution

Sketch or describe the distribution of individual users' daily comment counts when it is right-skewed. Mark the mean, median, and 95th percentile.

What This Part Should Cover Guidance

Mass near zero, long right tail, and nonnegative count support.
Typical ordering median < mean < p95 .
Why the mean is pulled right by heavy users.

Part 2 - Analyze Group Averages

Now repeatedly take many random user groups of equal size n and compute each group's average daily comments. Describe the distribution of these group averages.

What This Part Should Cover Guidance

Sampling distribution of the sample mean.
Mean remains equal to the population mean under random sampling.
Variance shrinks roughly by 1/n , and standard error shrinks by 1/sqrt(n) .
Distribution becomes more normal as n grows if CLT conditions are reasonable.

Part 3 - Compare Summary Statistics

How do the mean, median, and 95th percentile of the sampling distribution compare with those of the original per-user distribution?

What This Part Should Cover Guidance

Mean of sample means stays near the population mean.
Median of sample means approaches the mean as the sampling distribution becomes symmetric.
95th percentile of sample means is much closer to the mean than the original user-level p95.
Limitations when data are extremely heavy-tailed or users are not independent.

What a Strong Answer Covers Guidance

A strong answer distinguishes individual count distributions from sampling distributions, orders summary statistics correctly for right skew, and uses the CLT to explain why group averages are less variable and more symmetric.

Follow-up Questions Guidance

What if user comments follow a power-law distribution with extreme outliers?
How would excluding zero-comment users change the summary?
How would cluster sampling affect the variance of group averages?