Scribd
Upload
6 views

Formula_Cheat_Sheet

This cheat sheet provides essential formulas and concepts in statistics and econometrics, covering statistical distributions, point estimation, hypothesis testing, test statistics, linear regression, causal inference, panel data methods, and directed acyclic graphs. It includes key properties, decision rules, and methods for evaluating treatment effects. The document serves as a quick reference for statistical analysis and econometric modeling.

Uploaded by

elcanogaspar
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
6 views

Formula_Cheat_Sheet

This cheat sheet provides essential formulas and concepts in statistics and econometrics, covering statistical distributions, point estimation, hypothesis testing, test statistics, linear regression, causal inference, panel data methods, and directed acyclic graphs. It includes key properties, decision rules, and methods for evaluating treatment effects. The document serves as a quick reference for statistical analysis and econometric modeling.

Uploaded by

elcanogaspar
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Statistics & Econometrics Formula Cheat Sheet

Statistical Distributions
Distribution Formula Expectation & Variance

Bernoulli X ~ Ber(p) E(X) = p, Var(X) = p(1-p)

Binomial X ~ Bin(n, p) E(X) = np, Var(X) = np(1-p)

Poisson X ~ Pois(lambda) E(X) = lambda, Var(X) = lambda

Normal X ~ N(mu, sigma^2) E(X) = mu, Var(X) = sigma^2

Student t X ~ t(v) Symmetric, heavier tails

Chi-square X ~ chi^2(k) Sum of k squared normal variables

F-distribution X ~ F(df1, df2) Ratio of two chi-squares

Point Estimation & Inference


Concept Formula Key Property

Unbiasedness E(theta_hat) = theta Estimators mean equals true parameter

Efficiency Var(T) <= Var(T') Lower variance is preferred

Consistency lim E(Tn) = theta Estimator converges to theta

Maximum Likelihood L(theta) = Pi f(Xi|theta) Find theta that maximizes L(theta)

Hypothesis Testing
Type Null Hypothesis (H0) Alternative Hypothesis (H1)

One-Sample Mean mu = mu0 mu != mu0, mu > mu0, mu < mu0

One-Sample Proportion p = p0 p != p0, p > p0, p < p0

Two-Sample Mean mu1 = mu2 mu1 != mu2

Two-Sample Proportion p1 = p2 p1 != p2

P-Value & Critical Value


Test Decision Rule Significance Level

P-value Method p <= alpha -> Reject H0 alpha = 0.05 (default)


Critical Value |Test Stat| > Critical Value -> Reject H0 Z = +-1.96 for alpha=0.05

Test Statistics
Test Formula

Z-test Z = (X_bar - mu) / (sigma/sqrt(n))

T-test t = (X_bar - mu) / (s/sqrt(n))

F-test F = (S1^2 / S2^2)

Chi-square chi^2 = Sum [(O - E)^2 / E]

Likelihood vs Probability
Concept Definition

Probability Given parameters, what is the chance of data?

Likelihood Given data, what is the best parameter?

MLE theta_hat = argmax L(theta)

Linear Regression Model


Concept Formula Notes

OLS Estimator beta_hat = (X'X)^-1 X'Y Minimizes squared errors

R-Squared 1 - (SS_resid / SS_total) Explained variance

Homoskedasticity Var(epsilon) = constant Assumption MLR.5

Heteroskedasticity Test Breusch-Pagan BP Test in R

Causal Inference
Method Equation Purpose

Difference-in-Differences Delta_Y = (T_post - T_pre) - (C_post - C_pre) Policy evaluation

Instrumental Variables Z -> X -> Y, Cov(Z,epsilon) = 0 Corrects endogeneity

Regression Discontinuity Y = beta0 + beta1X + beta2D + epsilon Cutoff-based estimation

Panel Data Methods


Method Equation Use Case
Pooled OLS Yit = beta0 + beta1Xit + epsilon_it No individual effects

Fixed Effects Yit = beta1Xit + alpha_i + epsilon_it Controls for unit heterogeneity

Random Effects Yit = beta0 + beta1Xit + u_i + epsilon_it Assumes random variation

Directed Acyclic Graphs (DAGs)


Concept Definition

Backdoor Path Unblocked non-causal path

Collider D <- X1 -> Y

Confounder X1 -> D -> Y

Difference-in-Differences
Step Action

Step 1 Identify treatment & control groups

Step 2 Compare pre-treatment differences

Step 3 Compare post-treatment differences

Step 4 Compute treatment effect

You might also like