POLS 208 Lecture Notes - Lecture 12: Linear Regression, Null Hypothesis, Covariate
Document Summary
Regression is motivated by a desire to find the relationship between two variables Good for prediction and good for description (of those two variables) Estimating the average value of a variable given another variable. Linear regression looks for or estimates the line. Y = b +mx is drawn from . Price = a + b(weight) + ei. E reflects error in our attempt to relate weight to price. The best line is the one that minimizes the distance between the line and the points. The observed value of ei is the difference between the price we see and the price our line predicts given car weight. Srr (summing the square of those residuals) Weight on price /// weight (ind) price(dep) Weight on price catching on to the. So car prices are not only dependent on weight. Need to isolate/control for factors such as foreign cars .