POLS 208 Chapter Notes - Chapter 13: Regression Analysis, Piat, Confidence Interval
Document Summary
Regression can be used to estimate causal effects with only observational data. Can use more than one or two control variables at a time. And controls them at the same time. ** when x is exogenous, regression estimate causal effects. Wtatione usually need to add all the right exogenous variables. Multiple regression - predicts or explains a dependent variable using several independent variables. In multiple regression, the estimated effect of an independent variable of interest depends a great deal on which control variables are used. Examine the relationship between the independent variable of interest and potential common causes. Implement the control variable and observe the effect of the independent variable on the dependent variable, holding constant all the controls. The coefficient of a dummy variable is the difference in the average value of the dependent variable between those with and those without the characteristics.