Central Limit Theorem (Rigorous via Characteristic Functions)

Research Depth 91 in the knowledge graph I know this Set as goal
Unlocks 36 downstream topics
central-limit-theorem limit-theorems probability

Core Idea

If {Xₙ} are i.i.d. with mean μ and variance σ², then (Sₙ - nμ)/(σ√n) converges in distribution to N(0,1). The rigorous proof uses characteristic functions: φₙ(t/√n) → e^{-t²/2} for all t. The CLT explains why the normal distribution is ubiquitous—sums of many independent random variables are approximately normal regardless of the original distribution.

How It's Best Learned

Prove the CLT using characteristic functions. Apply the CLT to non-normal parent distributions to verify the approximation. Use the CLT to justify normal approximations in statistical inference.

Common Misconceptions

Explainer

You've studied characteristic functions — φ_X(t) = E[e^{itX}] — which encode all the distributional information about a random variable and behave nicely under sums (the characteristic function of a sum of independent variables is the product of their characteristic functions). You've also studied convergence in distribution, where a sequence of CDFs converges to a limiting CDF. The rigorous CLT proof combines these: it shows that the characteristic function of the standardized sum converges pointwise to e^{−t²/2}, the characteristic function of the standard normal, and then invokes the continuity theorem to conclude distributional convergence.

Here is the proof in outline. Without loss of generality, center and scale so that each Xᵢ has mean 0 and variance 1. The standardized sum is Sₙ/√n. Its characteristic function is φ_{Sₙ/√n}(t) = [φ_X(t/√n)]ⁿ. Now expand φ_X around 0: since E[X] = 0 and E[X²] = 1, the Taylor expansion gives φ_X(s) = 1 − s²/2 + o(s²). Substituting s = t/√n: φ_X(t/√n) = 1 − t²/(2n) + o(1/n). Raising to the n-th power: (1 − t²/(2n) + o(1/n))ⁿ → e^{−t²/2} for each fixed t. The continuity theorem then says: if the characteristic functions converge pointwise to the characteristic function of a distribution, the distributions converge in distribution. The standard normal has characteristic function e^{−t²/2}, so the result follows.

The finite variance condition is essential, not a mere technicality. The variance σ² appears as the coefficient in the Taylor expansion of φ_X: if variance is infinite, the second-order term is missing, the Taylor argument collapses, and the sum does not converge to a normal distribution. Instead, sums of heavy-tailed variables with infinite variance converge to stable distributions (of which the normal is a special case). The Cauchy distribution — whose variance is infinite — is the canonical example: sums of Cauchy variables rescaled by n give another Cauchy, not a normal. The CLT's universality is precisely bounded by the finite variance assumption.

Convergence in distribution — not almost sure, not in probability — is the correct mode here. The CLT says the *distribution* of (Sₙ − nμ)/(σ√n) approaches the standard normal. Individual observations remain drawn from whatever distribution they came from; what changes is the shape of the sampling distribution of the sum (or average). This is why "the CLT applies" in statistics means you can use normal critical values for large-sample inference: the sampling distribution of the sample mean is approximately normal. The approximation improves as n grows, but it is never exact for finite n (unless the original distribution is normal). Understanding this distinction — that the CLT is a statement about distributions, not individual outcomes — is the key to applying it correctly.

Practice Questions 5 questions

Prerequisite Chain

Counting to 10Counting to 20Understanding ZeroThe Number ZeroCounting to FiveOne-to-One CorrespondenceCombining Small Groups Within 5Addition Within 10Addition Within 20Two-Digit Addition Without RegroupingTwo-Digit Addition with RegroupingAddition Within 100Repeated Addition as MultiplicationMultiplication Facts Within 100Division as Equal SharingDivision as Grouping (Measurement Division)Division: Grouping (Repeated Subtraction) ModelDivision: Fair Sharing ModelDivision as Equal SharingDivision as GroupingBasic Division FactsDivision Facts Within 100Two-Digit by One-Digit DivisionDivision with RemaindersRemainders and Quotients in DivisionDivision Word ProblemsIntroduction to Long DivisionFactors and MultiplesPrime and Composite NumbersEquivalent FractionsRelating Fractions and DecimalsDecimal Place ValueReading and Writing DecimalsComparing and Ordering DecimalsAdding and Subtracting DecimalsMultiplying DecimalsDividing DecimalsDividing FractionsMixed Number ArithmeticOrder of OperationsInteger Order of OperationsVariable ExpressionsCombining Like TermsOne-Step EquationsTwo-Step EquationsSolving Multi-Step EquationsEquations with Variables on Both SidesAngle Pairs: Complementary, Supplementary, and VerticalParallel Lines and TransversalsCorresponding AnglesAlternate Interior AnglesTriangle Angle Sum TheoremExterior Angle TheoremTriangle Inequality TheoremSimilar Triangles: AA SimilaritySimilar Triangles: SSS and SAS SimilarityProportions in Similar TrianglesRight Triangle Trigonometry IntroductionTrigonometric Ratios ReviewRadian MeasureConverting Between Degrees and RadiansThe Unit CircleGraphing Sine and CosineGraphing Tangent and Reciprocal Trigonometric FunctionsDerivatives of Trigonometric FunctionsAntiderivativesIndefinite IntegralsBasic Integration RulesRiemann SumsDefinite Integral DefinitionFundamental Theorem of Calculus Part 1Fundamental Theorem of Calculus Part 2U-SubstitutionPartial Fraction Decomposition for IntegrationImproper Integrals - ConvergenceIntegral TestP-SeriesComparison TestLimit Comparison TestAbsolute vs. Conditional ConvergencePower SeriesTaylor PolynomialsTaylor SeriesMoment Generating FunctionsCharacteristic FunctionsConvergence in DistributionStationary DistributionsConvergence of Markov ChainsConvergence in ProbabilityAlmost Sure ConvergenceStrong Law of Large NumbersCentral Limit Theorem (Rigorous via Characteristic Functions)

Longest path: 92 steps · 510 total prerequisite topics

Prerequisites (4)

Leads To (3)