Chapter 3: Simple Linear Regression (Part 1)

3.1 The Simple Linear Regression Model 79
Using Temperature to Forecast Natural Gas Consumption
Time Series Data Treated as Cross Sectional
"Causal" model
Conditional mean, y intercept, slope, error

Using Value of House to Forecast Upkeep Expenditures

3.2 The Least Squares Point Estimates 88
Skim page 88 to the middle of  page 90
Point Prediction, Residuals, Sum of Squared Residuals
Least Squares Prediction Equation   (b versus beta)

3.3 Point Estimates and Point Predictions 92
Interpolation Within the Experimental Region
Extrapolation Outside the Experimental Region
Natural Gas Consumption when the temperature is 0 or 100

3.4 Model Assumptions and the Standard Error 96
Epsilon = the different potential effects on y of all factors other than x
Specifically, the way these factors move y above or below its conditional mean
Regression Assumptions: really assumptions about epsilon
Zero mean, Constant Variance, Normal Distribution, Independence
Be able to explain Figure 3.8 on page 98 to bosses, clients, politicians, news reporters
Mean Squared Error, Standard Error, Degrees of Freedom
Standard error estimates the standard deviation of y around the true conditional mean

x