Turn data into models,
one regression at a time.

One outcome, one explanatory variable, one line. Meet the population regression model, fit it by ordinary least squares, learn to read the slope as a marginal effect, measure how well the line fits with R², and reshape the relationship with logs and elasticities. Everything else in econometrics is this idea, scaled up.

1. 📐
   Level 1 The Regression Model ★★★ ◆◆
2. 🔒
   Level 2 Ordinary Least Squares ★★★ ◆◆ Needs: L01
3. 🔒
   Level 3 Slopes, Fits & Residuals ★★★ ◆◆ Needs: L02
4. 🔒
   Level 4 Goodness of Fit: R² ★★★ ◆◆ Needs: L03
5. 🔒
   Level 5 Logs, Elasticities & Functional Form ★★★ ◆◆◆ Needs: L04
6. 🔒 👑
   Level 6 · EXAM Simple Regression Exam ★★★ ◆◆◆ Needs: L05

Real questions have more than one cause. Add explanatory variables so each slope is a partial effect — the effect of one thing holding the others fixed. See how leaving a relevant variable out biases what's left (omitted-variable bias), fold categories in with dummy variables, and bend the model with interactions and quadratics.

1. 🔒
   Level 7 The Multiple Regression Model ★★★ ◆◆◆ Needs: L06
2. 🔒
   Level 8 Partialling Out: Ceteris Paribus ★★★ ◆◆◆ Needs: L07
3. 🔒
   Level 9 Omitted Variable Bias ★★★ ◆◆◆ Needs: L08
4. 🔒
   Level 10 Dummy Variables ★★★ ◆◆◆ Needs: L09
5. 🔒
   Level 11 Interactions & Quadratics ★★★ ◆◆◆ Needs: L10
6. 🔒 👑
   Level 12 · EXAM Multiple Regression Exam ★★★ ◆◆◆◆ Needs: L11

An estimate without a standard error is a guess with no error bars. Lay out the classical assumptions, see why OLS is unbiased and — under Gauss–Markov — the best you can do, then test claims: the t-test for one coefficient, the F-test for several, and what to do when heteroskedasticity threatens every standard error you just computed.

1. 🔒
   Level 13 The Classical Assumptions ★★★ ◆◆◆ Needs: L12
2. 🔒
   Level 14 Standard Errors & Gauss–Markov ★★★ ◆◆◆◆ Needs: L13
3. 🔒
   Level 15 The t Test & Confidence Intervals ★★★ ◆◆◆◆ Needs: L14
4. 🔒
   Level 16 The F Test ★★★ ◆◆◆◆ Needs: L15
5. 🔒
   Level 17 Heteroskedasticity ★★★ ◆◆◆◆ Needs: L16
6. 🔒 👑
   Level 18 · EXAM Inference Exam ★★★ ◆◆◆◆ Needs: L17

Data that arrives in order breaks the 'random sample' story. Learn to regress one time series on another without fooling yourself, handle trends and seasonality, spot the serial correlation that inflates your confidence, and build an autoregressive forecast you can actually check out of sample. Then put it all together and graduate the range.

1. 🔒
   Level 19 Time Series Regression ★★★ ◆◆◆ Needs: L18
2. 🔒
   Level 20 Trends & Seasonality ★★★ ◆◆◆◆ Needs: L19
3. 🔒
   Level 21 Serial Correlation ★★★ ◆◆◆◆ Needs: L20
4. 🔒
   Level 22 Forecasting ★★★ ◆◆◆◆ Needs: L21
5. 🔒
   Level 23 Building a Credible Model ★★★ ◆◆◆◆ Needs: L22
6. 🔒 👑
   Level 24 · EXAM The Final Exam ★★★ ◆◆◆◆◆ Needs: L23