BUS 215 Lecture Notes - Lecture 13: Nonlinear Regression, Stepwise Regression, Mortgage Loan
Document Summary
Sometimes we need to predict something that has only two possible values (a binary dependent variable). Verizon customer switch cell phone providers when the current contract expires, or remain with. Such research questions would seem to be candidates for regression modeling because firms would have many possible predictors (such as a customer"s age, gender, length of time as an existing customer, past transaction history, and so on). Y to be a number between 0 and 1, denoting the probability of the event of interest. Unfortunately, if you perform an ordinary least-squares regression with a binary (0 or 1) response variable, there will be complications. Your predicted y values could be greater than one or less than zero, which would be illogical. Another issue is that your regression errors will violate the assumptions of homoscedasticity (constant variance). As the predicted y values vary from . 50 (in either direction) the variance of the errors will decrease and approach zero.