Midterm Topics
EC 421: Introduction to Econometrics
Note: In general, you do not need to memorize proofs. Just understand the steps. I might ask how you get from one step to the next. I won’t ask you to write down a full proof.
Slide Set 1: Intro
- The goal of econometrics
- Regression notation
- Basic concept of causality
Slide Set 2: Review I
- Population vs. sample
- Parameters vs. sample estimates
- Estimators and uncertainty
- Uncertainty
- Standard errors
- Hypothesis testing
- t tests
- F tests
- Forming hypotheses
- critical value
- p-value
- Confidence intervals
- Linear regression and OLS
- “Best-fit” line
- Residuals
- SSE
- Estimators: bias and variance
- Statistical inference
- Variance (and standard error) of the OLS estimator
- Regressions with R’s
lm
function
Slide Set 3: Review II
- Simple and multiple linear regression
- Model fit
- R squared
- Overfitting
- Adjusted R squared
- Omitted-variable bias
- Interpreting coefficients
- Simple linear regression
- Multiple linear regression (ceterus paribus)
- Continuous explanatory variables
- Categorical explanatory variables
- Interactions
- Specifications
- Linear-linear
- Log-linear
- Log-log
- Inference vs. prediction
Slide Set 4: Heteroskedasticity
- The meaning of each of our assumptions/requirements
- Heteroskedasticity
- What it is
- What it looks like
- Consequences for OLS
- Tests for heteroskedasticity
- Goldfeld-Quandt test
- White test
- Chi-squared distribution
- Null and alternative hypotheses of each test
- Interpretations/conclusions for each
- Strengths and weaknesses of each test
Slide Set 5: Living with Heteroskedasticity
- Misspecification
- Weighted least squares
- Heteroskedasticity-robust standard errors
- Correlated disturbances and ‘clustering’
Slide Set 6: Asymptotics and Consistency
- Asymptotics
- Compared to ‘finite-sample’ attributes (probability limits vs. expected values)
- Probability limits
- Consistency
- Signing the bias from omitted variables.
- Measurement error and attenuation bias: What are they?
- Examples of measurement error