POL222H1 Lecture Notes - Lecture 8: Covariance, Statistical Inference, Linear Regression
Document Summary
Cannot randomly assign independent variable - can only observe, compare the average value of y across observations with different values of x. X and z may take many different values. Linear regression is arguably the most basic method to generalize the simple divide and compare method to control for confounding variables. X takes two values only: x = 0 or x = 1. Calculate the average of y for each value of x (conditional average) How y would vary on averages as x changes. Many more values of independent: x = 0, 1, 2, 3, . Ex: conditional average of y when x = 0, 1, etc. Then compare the different averages across different values of x. How does the value of y change on average. Independent: ideology: in politics, people sometimes talk of left and right. Divide our data by different values of x. Calculate the average value of y for each value of x (conditional average)