M1-004M1: Mathematics & Statistics FundamentalsProbability FoundationsMedium
⚡Flashcard Drill➔Mastery:

Explain Linearity of Expectation and Additivity of Variance, Highlighting Necessary Preconditions

🎯 Active Recall Self-Check (Draft First, Verify After)
Login-free practice tracker: recall core answer, audit pitfalls, and log review status.
Step 1 · Draft your concise answer & assumptions
Step 2 · Check Pitfalls & Follow-up Inquiries
Next: M1-005 →
📐 Mathematical Definition
E[aX+bY]=aE[X]+bE[Y],Var(X+Y)=Var(X)+Var(Y)+2Cov(X,Y)\mathbb{E}[aX+bY]=a\mathbb{E}[X]+b\mathbb{E}[Y],\qquad \mathrm{Var}(X+Y)=\mathrm{Var}(X)+\mathrm{Var}(Y)+2\mathrm{Cov}(X,Y)
⚡ One-Sentence Core Answer
Expectation is universally linear without any independence assumptions; variance additivity strictly requires zero pairwise covariance (uncorrelatedness).

📌 Key Takeaways

  • •
    Linearity of expectation holds universally across any arbitrary joint distribution, even under complete mutual dependence.
  • •
    Variance additivity Var(X+Y)=Var(X)+Var(Y)\text{Var}(X+Y)=\text{Var}(X)+\text{Var}(Y) requires Cov(X,Y)=0\text{Cov}(X, Y)=0.
  • •
    CUPED variance reduction in online experimentation fundamentally exploits non-zero covariance to cancel noise.

📐 Mathematical Derivations

By definition of discrete or continuous integration, E[aX+bY]=∫∫(ax+by)p(x,y)dxdy=a∫x(∫p(x,y)dy)dx+b∫y(∫p(x,y)dx)dy=aE[X]+bE[Y]E[aX + bY] = \int\int (ax + by)p(x, y)dxdy = a\int x(\int p(x,y)dy)dx + b\int y(\int p(x,y)dx)dy = aE[X] + bE[Y]. No independence or distributional constraints are needed. For variance, Var(X+Y)=E[((X+Y)−(E[X]+E[Y]))2]=Var(X)+Var(Y)+2Cov(X,Y)\text{Var}(X+Y) = E[((X+Y) - (E[X]+E[Y]))^2] = \text{Var}(X) + \text{Var}(Y) + 2\text{Cov}(X, Y). Hence, additivity holds if and only if Cov(X,Y)=0\text{Cov}(X, Y) = 0. If XX and YY are negatively correlated, the variance of their sum is strictly smaller than the sum of variances.

🏭 Production Trade-offs

In A/B testing and metric evaluation: (1) Sample means reliably estimate population means regardless of internal user interactions because expectation is unconditionally linear. (2) For ratio metrics like Click-Through Rate ∑Clicks∑Impressions\frac{\sum \text{Clicks}}{\sum \text{Impressions}}, numerator and denominator are strongly correlated across user clusters, so standard variance formulas fail and delta method or bootstrap must be used. (3) CUPED constructs Y∗=Y−θ(X−E[X])Y^* = Y - \theta (X - E[X]) with optimal θ=Cov(Y,X)Var(X)\theta = \frac{\text{Cov}(Y, X)}{\text{Var}(X)}, reducing variance to Var(Y)(1−ρ2)\text{Var}(Y)(1 - \rho^2).
⚠️ Common Interview Pitfalls
  • ✕
    Assuming variance is linear or additive when variables are correlated.
  • ✕
    Confusing 'uncorrelated' (Cov=0\text{Cov}=0) with 'statistically independent' (only equivalent for joint Gaussian distributions).
🎯 Interviewer Follow-up Inquiries
  • ?
    How does the Delta Method approximate the expectation and variance of non-linear transformations g(X)g(X)?
  • ?
    Why does cluster randomization require hierarchical variance estimation instead of simple i.i.d. variance additivity?
⚡
Test muscle memory on this topic via 3-Layer Flashcards
Active Recall: Mentally formulate Prompt ➔ Math Derivation ➔ Benchmark Implementation in 3 minutes.
🎯 Drill This Card Now➔
📚

Associated Knowledge Base Guides & Mindmaps

← PreviousM1-003: Define Independence vs. Conditional Independence, and Why Naive Bayes Works Despite Strong Violations📋 Back to BankNext →M1-005: Formulate Markov's and Chebyshev's Inequalities, and Contrast Their Tightness and Applications