← Meta Interview Insights

Meta·Data Scientist·Technical Phone Screen·Intermediate

Intermediate
Apr 2026

Summary

Short stats question at Meta, probably a data or analytics role screen. Nothing crazy, just one conceptual question but it can trip you up if you're not precise with the language.

Questions Asked (1)

Q1

What is the difference between covariance and correlation, and when would you use each?

Product Analytics & MetricsTechnical Trade-offs
Author's notes

Seemed basic until I started explaining it out loud.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by defining both concepts clearly, emphasizing that covariance measures the direction of a linear relationship while correlation measures both direction and strength on a standardized scale. Then explain the key mathematical difference (correlation is normalized covariance) and provide practical scenarios for when each is appropriate, tying it to data science work at Meta.

Pro tip: Mention that correlation is unitless and bounded between -1 and 1, making it easier to compare relationships across different scales, but also note that correlation only captures linear relationships and can be misleading with outliers or nonlinear patterns.

1. Define Covariance

Explain that covariance measures how two variables change together, indicating the direction of their linear relationship. It can range from negative to positive infinity and depends on the units of the variables.

2. Define Correlation

Explain that correlation is a standardized version of covariance, typically Pearson's r, which ranges from -1 to 1 and measures both the strength and direction of a linear relationship, independent of units.

3. Highlight Key Differences

Emphasize that correlation is scale-invariant and easier to interpret, while covariance is scale-dependent and useful for understanding raw joint variability. Correlation is covariance divided by the product of standard deviations.

4. When to Use Each

Discuss scenarios: use covariance when you need to understand the direction of relationship in original units or for mathematical derivations (e.g., portfolio variance). Use correlation for comparing relationships across different datasets or variables, and for feature selection in machine learning.

5. Connect to Data Science at Meta

Relate to product analytics: correlation is often used to identify metrics that move together (e.g., user engagement and revenue), while covariance might be used in multivariate analysis or A/B testing to understand joint variability.

Key Points to Mention

  • Covariance formula: cov(X,Y) = E[(X - E[X])(Y - E[Y])]
  • Correlation formula: corr(X,Y) = cov(X,Y) / (σ_X * σ_Y)
  • Correlation is unitless and bounded [-1, 1]; covariance is not
  • Correlation measures linear relationship only; can be misleading for nonlinear relationships
  • Use correlation for feature selection and comparing relationships across different scales
  • Use covariance in finance (portfolio theory) and for understanding joint variability in original units

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