SOC222H5 Lecture Notes - Lecture 9: Coefficient Of Determination, Regression Analysis
Document Summary
Regression models raised on line of fit. Least amount of erros square-ordinary least square. The more error there is a model, less of a good job. Lamda proportion amount of error without independant v with dependant v. Big chunk on regression y hat= a+bx. You would have to do regression for every value. How good we are doing in explaining an outcome. Smaller the r square we are doing a pretty bad job explaining the dependant variable. Education explains 59% change or variance for homosexuality. Economists think 0. 1 is enough to explain. R squared always increase by adding more variables. Regression model can have many independent variable. Write the line of fit-on lab and test.
